Cupola360 Insights & Stories

不是所有 360° 相機都一樣：如何選擇適合的 Camera Architecture？

Tue, 22 Sep 2026 08:46:34 GMT

市面上許多相機都能提供廣角、多方向，甚至 360° 的影像，但相似的觀看範圍，不代表它們使用相同的影像架構，也不代表它們適合相同的使用情境。

PTZ 著重遠距離目標與細節；多方向相機希望從一個安裝點同時監看數個重要方向；魚眼相機追求以簡單架構取得大範圍 overview；360° 運動相機重視完整拍攝、攜帶性與後製彈性；而拼接式全景相機則希望將多個方向整合成一個連續場景，讓完整環境可以被即時觀看、管理與分析。

因此，在進入解析度、FOV、FPS 或其他產品規格之前，更值得先問：

這種 Camera Architecture 原本是為了解決什麼問題？

沒有一種架構適合所有場景。真正重要的是先確認需要哪一種「看現場的方式」，再進一步比較產品規格。

選擇相機，應該先理解它是為了解決什麼問題

五種相機架構，其實在解決五種不同的視覺需求

Cupola360 的設計理念不是先決定「現在要看哪個方向」，而是先把完整場景保留下來。

多個影像感測器取得不同方向的畫面後，Cupola360 透過專用 panoramic image processing silicon，在相機內完成幾何校正、影像處理、projection 與即時硬體 stitching。也就是說，影像離開相機時，已經是完成拼接的連續 panoramic stream，而不需要先把多路影像送到後端，再由 server 或觀看軟體重新組成全景。相關晶片架構也包含 geometry stitching、luma / chroma correction 與 in-chip perspective projection 等處理能力。

這種架構讓不同方向不只是「同時被拍到」，而是存在於同一個具有空間關係的場景中。操作人員可以即時瀏覽完整環境，也可以從同一份全景資料選擇不同觀看方向、進行事件回顧，或將影像提供給其他系統使用。

Cupola360 的 stitched panoramic data 可依產品與專案需求整合 VMS、AI server、Cupola360⁺ Patrol、Robot platform 與其他上層應用，因此它的價值不只是提供一個 360° viewing experience，而是讓 panorama 本身成為可以被系統使用的 visual data。

為什麼機內硬體即時拼接對 AI 更重要？

360° 運動相機：完整拍攝，之後再決定怎麼呈現

Visual AI 不只是辨識畫面中的物件，更需要理解物件與周遭環境之間的關係。當 AI、機器人或遠端管理系統取得的是一份完整且連續的 panoramic scene，就能保留更多事件發生時的周邊脈絡，而不是只依賴事先選定的單一方向。

因此，對 Visual AI、Spatial Intelligence、Robot Vision，以及未來 Physical AI 等需要理解空間與真實環境的應用而言，完整全景資料可以作為更豐富的視覺資訊基礎。

Cupola360 的核心並不是單純「會 stitching」，而是透過 dedicated panoramic processing，在 edge 直接產生可持續輸出、可整合的 panoramic data。這也是內部技術比較中所整理出的主要架構差異。

360° 運動相機同樣希望在拍攝當下保留周遭更多方向的影像，但它的產品設計目標與企業全景相機不同。這類產品通常重視輕巧、穩定、防手震、便利操作與內容製作，使用者可以先把周遭拍下來，再於後製階段重新取景、調整觀看方向或輸出一般平面影片。

因此，它非常適合旅遊、運動、戶外活動、活動紀錄與內容創作。它與 Cupola360 最大的差異，不是誰「比較像 360° Camera」，而是使用 workflow 不同：360° 運動相機偏向完整拍攝、之後處理；Cupola360 則偏向完整拍攝、立即使用，並進一步支援長時間運行、遠端管理與企業系統整合。

魚眼相機：從單一位置取得大範圍 Overview

魚眼相機使用超廣角鏡頭，從單一位置取得非常大的觀看範圍，因此可以用較精簡的設備與安裝方式掌握室內空間。辦公室、零售、大廳、會議室與其他公共空間，都很常使用這類架構取得整體 overview。

魚眼原始影像會呈現明顯的曲面變形，因此通常需要透過 Dewarping 將部分區域轉換成較自然的視角。這種做法很適合快速掌握「這個區域大致發生什麼事」，但它的核心目的與多 sensor 拼接形成完整連續 panorama 並不相同。

換句話說，魚眼解決的是 如何用簡單架構看到更大的區域 ；拼接式全景則更進一步關心 如何保留完整場景及其空間關係 。

多方向相機通常把數顆獨立 sensor 整合在同一個機身中，每一個方向都可以依場域需求進行配置。因此，當使用者已經知道「哪些方向需要長期監看」時，這種架構可以有效減少相機機身數量、安裝點與佈線複雜度。例如路口、停車場、建築入口、倉庫、校園或大型公共空間，都可能利用一台多方向相機同步掌握數個重要區域。

但「多顆鏡頭」並不一定代表「一個連續 panorama」。這類產品常見的設計是讓每個方向維持獨立畫面，因此它解決的是 如何從同一個設備監看數個方向 ；Cupola360 所採用的拼接式架構則是 將不同方向整合成同一個連續場景 。兩種方法沒有絕對好壞，差異在於使用者需要的是「多個已知方向」，還是「完整空間脈絡」。

Supporting AI-Enabled Monitoring Across Different Sites

多方向相機：一個安裝點，同時看多個重要方向

PTZ：當任務是把遠方的重要目標看清楚

PTZ 的核心能力是 Pan、Tilt 與 Zoom。相機可以主動轉向、俯仰並利用光學變焦取得遠距離細節，因此特別適合交通、港口、機場、大型園區、周界、體育場與其他需要遠距觀察的環境。

當操作人員發現特定人物、車輛或事件時，PTZ 可以將注意力集中到目標並取得更清楚的細節，這是它非常重要的優勢。另一方面，它的設計邏輯本來就是「選擇一個重要方向來看」，因此完整場景掌握通常會搭配 overview camera 或其他影像來源。

PTZ 與全景相機因此不是彼此取代的關係。PTZ 解決的是 如何把特定目標看得更清楚 ；完整全景相機解決的是 如何不要失去其他方向正在發生的事情 。在實際系統中，兩種架構也可以互補使用。

Cupola360 即時全景相機：完整場景，而且立即可用

解析度、FOV 與 FPS 當然仍然是重要產品規格，但它們應該是在確認 Camera Architecture 之後才進一步比較的條件。

如果影像主要給人員看遠方的目標，光學 zoom 與 target tracking 可能更重要；如果只是希望以一個安裝點快速掌握室內區域，簡單的廣角 overview 可能已經足夠。但如果影像還需要進一步支援 AI 分析、遠端巡檢、Robot Vision、事件回顧或其他系統整合，那麼完整場景如何產生、是否能即時取得，以及後端能否直接使用這份影像資料，就會變得更加重要。

也因此，單純以 sensor 數量或最高解析度判斷哪一種 architecture「最好」，往往無法反映實際部署價值。Cupola360 內部的 competitive analysis 也指出，真正值得比較的還包括 stitching architecture、latency、完整 panoramic output、data accessibility，以及影像是否能直接提供給 AI、VMS、Robot 等系統。

比規格更重要的是：影像最後要怎麼被使用？

五種相機，其實是在回答五個不同問題

Cupola360 即時全景相機：我要持續掌握完整現場，而且影像可以立即被觀看、管理、分析與整合。
360° 運動相機：我要把周遭完整拍下來，再決定最後要如何呈現。
魚眼相機：我要從單一位置，以簡單方式取得大範圍 overview。
多方向相機：我已經知道數個重要方向，希望從一個安裝點同步監看。
PTZ：我要鎖定遠方的重要目標，並進一步取得細節。

不只選一種相機，而是建立更完整的視覺系統

在真實的企業場域中，往往不存在「只需要一種相機」的情況。

PTZ 可以負責遠距離細節，固定式相機可以持續監看重要區域，多方向相機適合預先規劃好的視角，而 360° 全景相機則可以提供更完整的現場脈絡。真正有效的影像系統，往往不是讓其中一種架構取代其他相機，而是讓不同視角彼此補充。

這也是 Cupola360 不只把自己定位成一台 360° 相機的原因。

Cupola360 的全景相機可以先提供完整場景視野；進一步搭配 Cupola360⁺ Patrol ，還能將全景影像、固定式 IP Camera、PTZ 與其他相容影像來源整合到同一個管理介面中。操作人員可以先從全景視角理解整個現場，再依需求查看特定區域或細節，不需要在彼此分離的監控畫面之間自行拼湊事件脈絡。

這使 Cupola360 的角色從單一 Camera Architecture，進一步延伸成一個更完整的 panoramic visual layer ：不只是拍攝完整現場，也讓不同影像來源可以被組織、觀看，並進一步提供給 AI、遠端巡檢與其他營運系統使用。Cupola360 的開放式架構本身也以讓 panoramic data 能進入 VMS、AI server、Robot 等上層系統為重要設計方向。

選擇相機時，沒有必要把 PTZ、魚眼、多方向相機與全景相機看成只能「五選一」的競爭關係。不同架構本來就擅長不同工作，而一個完整的企業影像系統，往往需要不同視角彼此配合。

Cupola360 的核心價值，在於補上其中最容易缺少的一塊： 完整的場景脈絡。

透過晶片內即時硬體拼接，Cupola360 在相機端就能將多個 sensor 的影像轉換成連續的 panoramic data，讓完整 360° 場景可以立即被觀看、管理、分析與整合。這份完整場景資訊可以進一步支援 Visual AI、Spatial Intelligence、Robot Vision 與未來 Physical AI ，讓 AI 不只看見單一物件，也能取得更多物件與周遭環境之間的空間關係。

而當專案中同時存在 PTZ、固定式 IP Camera 或其他影像設備時，Cupola360 的價值也不必停留在「多裝一台全景相機」。透過 Cupola360⁺ Patrol ，不同影像來源可以進一步被整合到同一個管理視角，讓使用者從完整場景出發，再深入需要關注的局部細節。

所以 Cupola360 真正提供的，不只是一個全景相機，而是 以完整全景資料支援 AI，並以統一視覺介面串起遠端場域管理。

結論：相機架構不同，但現場需要的是完整視野

Not All 360° Cameras Are Built the Same: Choosing the Right Camera Architecture

Tue, 22 Sep 2026 08:36:41 GMT

Many cameras can provide wide-angle, multidirectional, or even 360° imagery, but similar coverage does not mean they use the same architecture or serve the same application.

PTZ cameras prioritize long-range targets and detail. Multidirectional cameras monitor several important directions from one installation point. Fisheye cameras provide a broad overview with a relatively simple optical design. 360° action cameras emphasize immersive capture, portability, and post-production flexibility. Stitched panoramic cameras address another requirement: preserving multiple directions as one continuous scene that can be viewed, managed, and analyzed in real time.

Before comparing resolution, FOV, frame rate, or individual specifications, it is therefore more useful to ask:

What problem was this camera architecture designed to solve?

There is no single architecture for every application. The right starting point is the way the scene needs to be seen and used.

Start With the Problem the Camera Was Designed to Solve

Cupola360 is designed around a different principle: instead of deciding which direction matters first, capture and preserve the complete scene.

Multiple image sensors capture different directions, while dedicated panoramic image processing silicon performs geometric correction, image processing, projection, and real-time hardware stitching inside the camera. The output leaving the camera is therefore already a continuous stitched panoramic stream rather than a set of separate views that must first be reconstructed by a downstream server or viewing application. The underlying processing architecture includes geometry stitching, luminance and chroma correction, and in-chip perspective projection.

The result is more than simultaneous coverage of several directions. Those directions remain part of one continuous spatial scene. Operators can navigate the environment in real time, select different viewpoints from the same panoramic data, review events, and make the video available to other systems.

Depending on the product and deployment architecture, Cupola360 panoramic data can be integrated with VMS platforms, AI servers, Cupola360⁺ Patrol, robotic platforms, and other operational applications. The panorama is therefore not limited to a viewing experience—it can become usable visual data for a broader system.

Five Architectures for Five Different Visual Needs

Visual AI increasingly needs more than isolated object detection. Many applications also need context: where an object is located, what surrounds it, and how it relates to the wider physical environment.

A continuous panoramic scene preserves more of this spatial context than a single preselected direction. For Visual AI, Spatial Intelligence, Robot Vision, and future Physical AI applications , this broader scene can provide a richer visual foundation for detection, navigation, event understanding, and interaction with physical environments.

Cupola360's differentiation is therefore not simply that it can stitch video. The architecture uses dedicated panoramic processing at the edge to generate continuous panoramic data that is immediately available for downstream systems. This is also the more meaningful architectural distinction identified in Cupola360's competitive analysis.

360° Action Cameras: Capture First, Decide the View Later

Fisheye Cameras: Broad Coverage From a Single Viewpoint

360° action cameras also aim to preserve imagery from multiple directions, but their workflow is generally built around mobility, stabilized capture, immersive viewing, and content creation. Users can record the surrounding environment first and later choose a viewing direction, reframe the scene, or export a conventional video during post-production.

This makes the architecture highly suitable for travel, sports, outdoor activities, events, and personal content. The important difference is not whether both products are called “360° cameras.” It is how the video is intended to be used: action cameras are typically capture first, process later , while Cupola360 is designed around capture the complete scene and use it immediately in continuous monitoring and integrated operational workflows.

Fisheye cameras use an ultra-wide lens to capture a large area from one installation point. Their compact architecture makes them useful for offices, retail spaces, lobbies, meeting rooms, and other indoor environments where a broad overview is more important than long-range detail.

Because the original fisheye image is heavily curved, software dewarping is typically used to create more natural viewing perspectives. This is effective for quickly understanding the general condition of an area, but the architecture is fundamentally different from combining multiple sensors into one stitched panoramic scene.

A fisheye camera is primarily about seeing more area from one point. A stitched panoramic architecture focuses on preserving a complete scene and its spatial relationships.

Multidirectional Cameras: Several Important Directions From One Device

A multidirectional camera combines multiple independent sensors within one housing. Each sensor can be positioned toward a different area, making the architecture useful when the important viewing directions are already known. Intersections, parking areas, entrances, warehouses, campuses, and other multi-zone environments can benefit from fewer camera bodies, installation points, and cables.

However, multiple sensors do not automatically create one continuous panorama. In many multidirectional architectures, each direction remains an independent view. The system solves the pro blem of monitoring several important directions from one device, while a stitched panoramic architecture focuses on connecting those directions into one continuous spatial scene.

Neither approach is universally better; they address different operational needs.

PTZ cameras are designed to pan, tilt, and optically zoom toward important targets. This makes them highly effective for traffic, ports, airports, large campuses, perimeters, stadiums, and other environments where operators need to identify or follow distant activity.

Their strength comes from concentrating visual attention on a selected target and capturing greater detail. Because the architecture actively chooses a viewing direction, wider situational awareness is often complemented by overview cameras or other video sources.

PTZ and panoramic cameras therefore serve different purposes. PTZ answers “How can I see this target more clearly?” Panoramic vision answers “How can I preserve what is happening across the entire scene?” In many systems, the two approaches can also work together.

Supporting AI-Enabled Monitoring Across Different Sites

PTZ Cameras: When the Priority Is Long-Range Detail

More Important Than Specifications: How Will the Video Be Used?

Cupola360: Complete Scene Data, Ready in Real Time

Why In-Camera Hardware Stitching Matters for AI

The important distinction is not simply whether a camera can “see 360°.” It is what kind of visual data comes out of the architecture and what the rest of the system can do with it .

Five Architectures, Five Different Questions

Beyond One Camera Type: Building a More Complete Visual System

Conclusion: Different Cameras, One More Complete View

Resolution, field of view, and frame rate remain important product specifications, but they become much more meaningful after the application architecture is clear.

If the goal is long-range target observation, optical zoom and tracking may matter most. If the requirement is a simple indoor overview, a broad single-point view may be sufficient. But when video also needs to support AI analytics, remote inspection, robot vision, incident review, or system integration, questions such as how the complete scene is generated, whether it is available in real time, and whether downstream systems can directly use the data become increasingly important.

For this reason, sensor count or maximum resolution alone cannot determine which architecture is “better.” Cupola360's internal competitive analysis similarly identifies stitching architecture, latency, panoramic output, and data accessibility to AI, VMS, robots, and other platforms as important architectural considerations.

Cupola360: I need continuous awareness of the complete scene, with video that can be viewed, managed, analyzed, and integrated immediately.
360° Action Camera: I want to capture everything around me and decide later how the content should be presented.
Fisheye Camera: I want a simple way to get a broad overview from one installation point.
Multidirectional Camera: I already know the important directions and want to monitor them simultaneously from one device.
PTZ: I need to focus on a distant target and see it in greater detail.

In real-world enterprise environments, there is rarely a need to rely on only one camera architecture.

PTZ cameras can provide long-range detail. Fixed IP cameras can continuously monitor critical areas. Multidirectional cameras can cover several predefined views, while panoramic cameras provide broader scene context. A stronger visual system does not necessarily replace one architecture with another—it allows different viewpoints to complement each other.

This is why Cupola360 is designed to go beyond the role of a standalone 360° camera.

Cupola360 panoramic cameras provide the full-scene visual layer. With Cupola360⁺ Patrol , panoramic video, fixed IP cameras, PTZ cameras, and other compatible video sources can be brought into a unified management interface. Operators can begin with the broader panoramic context and then move into specific areas or detailed views without having to reconstruct the event mentally across disconnected video feeds.

This extends Cupola360 from a single camera architecture into a broader panoramic visual layer —one that not only captures the complete scene, but also helps organize, view, and deliver visual data to AI, remote inspection, and other operational systems. Cupola360’s architecture is also designed to make panoramic data accessible to VMS, AI servers, robotic platforms, and other upper-layer solutions.

PTZ, fisheye, multidirectional, and panoramic cameras are not necessarily mutually exclusive choices. Each architecture is designed for a different visual task, and enterprise environments often benefit from combining them.

Cupola360 adds something that is especially valuable in this mix: complete scene context.

With real-time on-chip hardware stitching, multiple sensor inputs are transformed inside the camera into continuous panoramic data that c an be viewed, managed, analyzed, and integrated immediately. This richer scene context can support Visual AI, Spatial Intelligence, Robot Vision, and future Physical AI applications, giving downstream systems more information about how objects relate to the surrounding environment.

When PTZ, fixed IP cameras, or other video sources are also part of the deployment, Cupola360 does not have to remain another isolated camera. Through Cupola360⁺ Patro l, multiple video sources can be brought together into a unified visual experience—allowing operators to understand the complete scene first, then move into the detail that matters.

Cupola360 is therefore more than a 360° camera. It delivers complete panoramic data for AI and a unified visual layer for remote site management.

從全景視覺到數位孿生：打造 AI 資料中心的遠端維運能力

Fri, 18 Sep 2026 06:24:37 GMT

將 360° 全景影像、設備端感測、熱資訊與電力資料帶進數位孿生，讓資料中心的實體環境變得可視、可判讀，也更容易遠端管理。

隨著 AI 運算需求快速成長，資料中心正朝向更高運算密度、更高機櫃功耗，以及更複雜的供電與散熱架構發展。

傳統伺服器管理系統已能持續掌握溫度、電壓、風扇與系統健康狀態，但當單一機櫃承載更多 GPU、電力設備與冷卻系統後，影響營運穩定性的訊號，往往不只存在於伺服器內部。

維運團隊還需要了解：

PDU、UPS 與電力節點目前的狀態
機櫃或設備是否出現局部異常升溫
液冷管線與周邊環境是否有異常
機櫃門板是否開啟、設備位置是否改變
設備燈號、QR Code 與資產標籤是否與系統資料一致
異常發生時，周邊現場究竟發生了什麼

這些資訊往往散落在不同設備、畫面與管理系統中，其中不少狀態甚至仍需要人員親自到場確認。當資料中心規模擴大、機櫃密度提高，單純增加人工巡檢不只成本更高，也很難即時掌握快速變化的異常。

先讓資料中心的「實體世界」被看見

數位孿生要真正協助資料中心維運，不能只有設備清單與數值。管理者還需要知道：這個設備實際位在哪裡？周邊環境是什麼？異常發生時，現場看起來如何？

Cupola360 在這套合作方案中的角色，就是提供場域級 360° 全景視覺。透過全景相機，維運人員可以從較完整的空間視角掌握機房環境、機櫃配置與周邊狀態；當某個區域發生異常，也能從完整現場脈絡理解事件位置及周邊設備，而不是只看到單一固定鏡頭。

方案同時結合巽晨國際的設備端感測能力，透過微型感測相機、熱影像與近距離設備視角，進一步查看 PDU、UPS、設備燈號、資產標籤與局部熱點。

因此，維運人員可以從： 看整個機房 → 找到特定機櫃 → 查看設備細節， 逐步深入，而不必在缺乏空間關係的多套系統中來回切換。

把影像、設備狀態與電力資料帶進數位孿生

套方案的重點，不是單純在資料中心增加更多攝影機。真正的價值，是把來自不同設備與感測器的 Physical Data，重新對應到真實的空間與設備位置。

數位孿生介面可以將真實世界中的房間、機櫃、UPS、PDU、感測設備與攝影機建立位置關係，並進一步整合：

360° 全景影像
設備端近距離影像
熱資訊
電力資料
冷卻資訊
設備告警
歷史運行資料

維運人員因此不再只看到一串數值，而能將「設備狀態」與「真實位置」放在同一個空間脈絡中理解。例如，當系統出現設備過熱告警時，操作人員可以由數位孿生中的設備位置，進一步切換至對應的實際全景畫面、設備近距離影像及歷史趨勢。

原本需要：

收到警報 → 找設備 → 到現場 → 查看狀況

才能完成的確認流程，就有機會轉變為：

收到警報 → 定位設備 → 查看現場 → 比對資料

縮短從「知道有問題」到「理解發生什麼事」之間的距離。

三方能力如何形成完整的維運視角？

從「看見異常」走向「理解異常」

數位孿生的價值，不只是「做一個 3D 機房」

不取代既有管理系統，而是補上 Physical Visibility

從 Physical Data 建立更主動的資料中心維運

關鍵亮點

以 360° 全景視覺建立資料中心的場域級 Physical Visibility
結合設備端微型感測相機與熱影像，查看機櫃與設備細節
將影像、電力、冷卻、環境與設備資訊對應至數位孿生
從告警快速連結設備位置、現場影像與歷史趨勢
支援遠端巡檢、異常定位、多來源交叉判讀與事件追溯
整合長期營運資料，支援能源、容量與趨勢分析
為後續 AI 異常預警與預測性維護建立資料基礎
補足 BMC、DCIM、EPMS 等既有管理系統較難直接取得的實體現場資訊

AI 資料中心的維運挑戰，已不只存在於伺服器內部

AI Data Center Digital Twin | Panoramic Vision and Remote Operations

Fri, 18 Sep 2026 04:11:14 GMT

Connecting panoramic vision and physical data through a digital twin for smarter data center operations.

As AI computing continues to increase rack density and power demand, data centers are moving toward more complex power distribution, cooling, and infrastructure architectures.

Traditional server management systems already provide detailed information on temperature, voltage, fans, and system health. But as individual racks carry more GPUs, power equipment, and cooling infrastructure, many operational signals now exist outside the server itself.

Operation teams may also need to understand:

The condition of PDUs, UPS systems, and electrical nodes
Local temperature increases around racks or equipment
Potential abnormalities around liquid cooling infrastructure
Rack door status and equipment position changes
Equipment indicators, QR codes, and asset labels
What was physically happening around an equipment alert

Much of this information remains distributed across different devices, interfaces, and management systems. Some conditions still require personnel to physically inspect the site. As data centers become larger and denser, relying only on additional manual inspection becomes increasingly difficult to scale.

Making the Physical Data Center Visible

For a digital twin to support real data center operations, equipment lists and numerical data alone are not enough.

Operators also need to understand where an asset is physically located, what surrounds it, and what the actual environment looks like when an abnormal condition occurs.

Cupola360 provides the site-level panoramic visual layer within this collaborative solution.

With 360° panoramic cameras, operation teams can view the broader data hall, rack layout, and surrounding environment within a more complete spatial context. When an abnormal condition occurs, operators can understand where the event is taking place and what is happening around it instead of relying only on a single fixed camera view.

The solution can also incorporate equipment-level sensing and close-range imaging for PDUs, UPS systems, status indicators, asset labels, and localized thermal conditions.

This creates a visual workflow that moves naturally from: the room → the rack → the equipment while maintaining the spatial relationship between each level.

Bringing Video, Equipment Status, and Power Data into the Digital Twin

The goal is not simply to add more cameras to the data center.

The greater value comes from connecting different sources of Physical Data to their real-world locations.

A digital twin can map rooms, racks, UPS systems, PDUs, sensors, and cameras to the physical environment and connect them with:

360° panoramic video
Equipment-level visual feeds
Thermal information
Power data
Cooling information
Equipment alarms
Historical operating data

This allows operators to understand equipment status in its real spatial context rather than as an isolated value on a dashboard.

When an overheating alert appears, for example, the operator can move from the equipment location in the digital twin to the corresponding panoramic view, close-range equipment information, thermal condition, and historical trend.

The workflow can therefore shift from:

Alert → Locate → Visit the site → Inspect

toward:

Alert → Locate → View the site → Compare the data

reducing the gap between knowing that something is wrong and understanding what is actually happening.

How the Partner Technologies Work Together

From Seeing an Alert to Understanding It

A Digital Twin Is More Than a 3D Data Center Model

Complementing Existing Management Systems with Physical Visibility

Turning Physical Data into More Proactive Data Center Operations

Related Resources

The solution combines different technology layers, each contributing a different part of the operational picture.

Cupola360 | Site-Level Panoramic Vision
Cupola360 provides 360° panoramic visual coverage of the data hall, rack layout, and surrounding environment.

Compared with isolated fixed-camera views, panoramic vision preserves broader spatial context and helps operators understand where an abnormal condition occurs and what surrounds it.

Millitronic｜Equipment-Level Vision and Sensing
Close-range sensing and imaging provide detailed visibility into racks, PDUs, UPS systems, equipment indicators, asset labels, and localized thermal conditions.

Together with panoramic vision, this allows the solution to support both site-level awareness and equipment-level inspection.

Schneider Electric | Infrastructure, Energy, and Long-Term Operational Data
Schneider Electric contributes data center infrastructure, power, and energy management capabilities that connect physical information with existing power, cooling, capacity, and equipment management data.

By accumulating historical information and analyzing long-term trends, operation teams can monitor changes in power consumption, temperature, cooling efficiency, capacity utilization, and equipment condition, helping them identify conditions that begin to deviate from normal operating baselines.

Remote operations should not simply move more screens into a control room. A useful operational workflow should help teams move from detection to location, verification, and review.

When panoramic video, equipment imagery, thermal information, power data, and historical trends are connected, the system can support:

Remote Inspection

Review data halls, racks, and key equipment without visiting each location in person.

Abnormality Localization

Move directly from an alert to the corresponding equipment, physical location, and visual context.

Cross-Source Verification

Compare status indicators, temperature, power values, and live visual information together.

Event Review

Use historical data and video to understand how conditions changed before and after an event.

Early Warning and Predictive Maintenance

As historical data accumulates, AI analysis can gradually support trend-based warning for thermal, electrical, and equipment abnormalities and provide a foundation for predictive maintenance.

The goal is not to replace data center operators with AI, but to give them more complete physical and historical context when decisions are required.

Digital twin discussions can easily become focused on 3D visualization itself.

But an operational digital twin creates more value when the virtual environment remains connected to real equipment, live information, and physical site conditions.

Selecting a rack should not only show where it appears in a 3D model. It should also connect the operator to relevant power status, temperature, equipment information, alarm history, and the actual visual environment around that rack.

This transforms the digital twin from a visualization tool into an operational interface for the physical data center.

This solution is not intended to replace BMC, DCIM, EPMS, or other existing server and infrastructure management platforms.

Instead, it adds something these systems may not directly provide: Physical Visibility.

What does the equipment actually look like right now?
Where is it physically located?
What is happening around it?
Can a numerical abnormality be verified through the actual site view?

A more precise way to describe the solution is therefore:

Extending data center visibility and operational management from inside the server to the entire physical environment.

As AI data centers move toward higher power density, greater computing capacity, and more complex cooling and power architectures, operational risk increasingly extends beyond server health.

Racks, power infrastructure, thermal conditions, liquid cooling, equipment indicators, and the surrounding physical environment are all becoming part of the operational data set.

By combining Cupola360 panoramic vision, equipment-level sensing, and Schneider Electric’s infrastructure, energy, and long-term operational data capabilities, distributed Physical Data can be transformed into information that is more visible, traceable, and easier to interpret.

The next step for data center operations is not simply connecting more equipment to software. It is making what happens in the physical world part of the digital operational workflow. .

Key Highlights

Establishes site-level Physical Visibility through 360° panoramic vision
Combines equipment-level sensing and thermal information for detailed rack and device inspection
Maps visual, power, cooling, environmental, and equipment information into a digital twin
Connects alerts with equipment location, live visual context, and historical trends
Supports remote inspection, abnormality localization, cross-source verification, and event review
Uses long-term operational data to support energy, capacity, and trend analysis
Creates a data foundation for future AI-based early warning and predictive maintenance
Complements BMC, DCIM, and EPMS platforms with physical site context

AI Data Center Operations Go Beyond the Server

高風險任務中的全景視野：360° 影像如何支援無人載具遠端勘查

Tue, 15 Sep 2026 08:37:53 GMT

透過 360° 全景影像，讓無人載具、機器狗與移動式巡檢平台在複雜場域中取得更完整的周邊視野。

在海軍陸戰隊成軍 79 週年隊慶活動中，Cupola360 與合作夥伴 LAKI ROBOTICS 展示了 360° 全景視覺在無人載具與移動式巡檢平台上的應用可能。

隨著無人地面載具、機器狗與自主移動平台逐漸被應用於基地巡檢、關鍵設施管理、災害勘查與高風險場域作業，遠端操作人員需要的不只是「看見前方」，而是能夠更完整理解載具周遭的現場狀態。

傳統前視鏡頭可以協助載具前進，但當任務環境變得更複雜時，單一方向的視野往往不夠。載具左右兩側是否有人員或障礙物？後方是否出現變化？周邊環境是否有短暫出現但很快消失的異常？這些資訊如果沒有被同步保留下來，就可能形成遠端操作與任務判斷上的盲區。

Cupola360 的 360° 全景相機可安裝於無人載具、機器狗、巡檢機器人或其他移動平台上，持續擷取平台周圍的完整影像。即使載具正在向前移動，遠端操作人員仍可同步查看前方、左右與後方環境，降低傳統固定視角或前視鏡頭可能造成的視覺死角。

這項展示的重點，不是把無人系統變得更複雜，而是讓移動平台在執行任務時，能夠帶回更完整、更有現場脈絡的視覺資訊。

無人系統的價值，不只是代替人員移動到遠方或危險區域，更重要的是讓人員能在安全距離之外，掌握現場、理解風險，並更快做出判斷。

但在實際應用中，無人載具常常面對的是不確定的環境：狹窄通道、基地周界、倉儲區、地下空間、受損建築、災害現場或其他高風險區域。這些地方的環境變化不一定只發生在載具正前方，也可能出現在側邊、後方或視線邊緣。

如果操作人員只能依賴前視鏡頭，就必須不斷轉動鏡頭、調整車身，或等待載具改變方向，才能確認周邊情況。這不只增加操作負擔，也可能讓短暫出現的異常被錯過。

360° 全景視覺的價值，正是在於讓無人系統不只「朝前看」，而是能夠保留完整周邊畫面。對遠端操作人員來說，這代表他們可以在同一個畫面中理解載具與環境的關係，而不是被單一鏡頭方向限制。

為什麼無人系統需要 360° 視野？

高風險區域勘查：先看清楚，再決定下一步

在高風險區域中，現場可能存在可疑物體、障礙物、結構損壞、煙霧、化學污染或其他未知風險。若直接派人進入，不僅需要更高的安全準備，也可能讓人員暴露在不必要的危險之中。

搭載 Cupola360 360° 全景相機的無人載具，可以先進入指定區域，將周邊道路、出入口、牆面、障礙物與環境變化即時傳回後方。遠端人員可以依照需要查看不同方向，不必完全依賴載具當下的行進方向。

這讓勘查不只是取得一段前方影像，而是取得一個完整的現場脈絡。當團隊需要評估是否派人進入、選擇哪條路線、避開哪些風險，或確認現場是否仍有變化時，360° 全景影像能提供更完整的判斷基礎。

基地與關鍵設施巡檢：讓巡檢紀錄保留空間脈絡

在基地、工廠、能源設施、倉儲、機庫、通信站或其他關鍵場域中，巡檢工作往往不只是在固定點拍攝畫面，而是需要理解整個場域的連續狀態。

無人載具可依照設定路線巡視周界、通道、設備區與限制區域。若只使用單一方向鏡頭，巡檢紀錄可能只保留載具前方畫面；但實際事件可能發生在側邊、後方或載具通過後的某個位置。

360° 全景影像可以讓巡檢紀錄更具空間脈絡。當系統發現異常，或後續需要回顧事件時，操作人員可以查看事件發生前後的完整周邊畫面，更容易判斷異常位置、移動方向、周邊人員或設備變化。

對需要長時間巡檢、多點管理或跨區域監看的場域而言，這種完整視覺紀錄能讓遠端管理更有效率，也讓事件追溯更有依據。

在移動式任務中，現場資訊通常不只提供給單一操作員。後方管理中心、安全單位、工程團隊或決策人員，都可能需要即時理解現場狀況。

傳統影像串流多半由前線操作員決定鏡頭方向，其他觀看者只能看到同一個角度。當多個角色關心的資訊不同時，單一視角就會限制協作效率。

360° 全景影像則可以讓不同使用者在同一個現場影像中，各自選擇觀看方向。操作員可以專注於載具移動，後方人員可以查看側邊或後方環境，工程人員可以觀察設備區，管理者也能快速掌握整體現場。

這讓遠端協作不再只是「看同一支鏡頭」，而是共同進入同一個全景場景，各自取得任務所需的資訊。

Cupola360 如何支援這項應用

在這類移動式場域感知應用中，Cupola360 提供的是 360° 全景視覺能力，協助合作夥伴的無人載具、機器狗或自主移動平台取得更完整的周邊影像。

相較於只依賴前視鏡頭，Cupola360 360° 全景相機能讓遠端操作人員在平台移動時同步查看前方、側邊與後方環境。這有助於降低視覺死角，並讓巡檢、勘查與遠端監看保留更完整的現場脈絡。

Cupola360 的價值不是取代無人載具或機器人平台，而是作為移動式系統上的全景視覺層，讓合作夥伴與系統整合商能在更完整的影像基礎上，發展巡檢、偵察、遠端監控、事件回顧與場域管理應用。

從移動平台到 Reality Remote Management

無人載具的發展，正在讓更多高風險、遠距離或人員不易長時間停留的場域，具備新的遠端管理方式。

但真正有價值的遠端管理，不只是把畫面傳回來，而是讓人員能理解現場、判斷事件，並進一步採取行動。這也是 Cupola360 所強調的 Reality Remote Management：不只看見現場，更能遠端理解、管理與應對。

當 360° 全景視覺與無人系統結合，移動平台就不只是前進的載具，也能成為持續收集現場脈絡的視覺節點。這讓巡檢、勘查、遠端協作與事件追溯，都能建立在更完整的現場資訊之上。

遠端態勢感知：讓不同單位進入同一個全景場景

應用亮點

讓無人載具、機器狗與移動式巡檢平台取得 360° 周邊視野
降低傳統前視鏡頭可能造成的視覺死角
支援高風險區域勘查、基地巡檢、關鍵設施管理與災害現場確認
協助遠端操作人員同步查看前方、左右與後方環境
保留完整現場脈絡，支援事件回顧與異常判斷
讓不同單位可在同一個全景場景中各自選擇觀看方向
可與合作夥伴無人載具、機器人平台與系統整合方案搭配應用

備註 : 本文介紹之內容為活動展示與應用情境說明。實際功能、防護規格、通訊方式、資料安全、離線儲存與環境耐受能力，仍需依專案需求、選用型號與系統整合方式確認。

A Wider View for High-Risk Missions: 360° Vision for Unmanned Field Inspection

Tue, 15 Sep 2026 08:31:27 GMT

Using panoramic cameras and AI detection to monitor outdoor areas, reduce blind spots, and respond with better context.

At the Marine Corps 79th Anniversary event in Taiwan, Cupola360 and partner LAKI ROBOTICS presented how 360° panoramic vision can support unmanned vehicles and mobile inspection platforms.

As unmanned ground vehicles, robotic dogs, and autonomous mobile platforms are increasingly explored for facility patrol, critical infrastructure monitoring, disaster inspection, and high-risk field operations, remote operators need more than a forward-facing camera. They need a broader understanding of the environment around the platform.

A traditional front-facing camera can help a vehicle move forward, but in a complex environment, a single viewing direction is often not enough. Are there people or obstacles on the sides? Has something changed behind the vehicle? Did a short-lived event appear at the edge of the scene and disappear before the camera turned?

Without full surrounding visibility, these details can become blind spots in remote operation and field decision-making.

Cupola360 360° panoramic cameras can be mounted on unmanned vehicles, robotic dogs, inspection robots, or other mobile platforms to continuously capture full surrounding video. Even while the platform is moving forward, remote operators can still view the front, sides, and rear at the same time, reducing visual blind spots created by traditional fixed-view or front-facing cameras.

The value of this application is not to make unmanned systems more complex. It is to help mobile platforms bring back more complete visual context from the field.

The value of unmanned systems is not only that they can move into remote or high-risk areas. Their real value is that they allow personnel to stay at a safer distance while still understanding the field, assessing risk, and making faster decisions.

In real deployments, unmanned platforms often operate in uncertain environments: narrow corridors, facility perimeters, warehouses, underground spaces, damaged buildings, disaster sites, or other high-risk areas. Important changes may not appear directly in front of the vehicle. They may occur on the side, behind the platform, or at the edge of a camera’s field of view.

If operators rely only on a front-facing camera, they may need to constantly rotate the camera, adjust the vehicle direction, or wait until the platform turns before they can confirm the surrounding situation. This increases operating effort and may cause short-lived events to be missed.

360° panoramic vision helps unmanned systems move beyond “looking forward.” It preserves a full surrounding view, allowing remote operators to understand the relationship between the platform and its environment within one visual context.

Why Unmanned Systems Need 360° Visibility

High-Risk Area Inspection: See the Scene Before Taking the Next Step

In high-risk areas, the site may include suspicious objects, obstacles, damaged structures, smoke, chemical contamination, or other unknown hazards. Sending personnel directly into these environments may require extensive safety preparation and could expose people to unnecessary risk.

An unmanned vehicle equipped with a Cupola360 360° panoramic camera can enter the area first and transmit a broader view of roads, entrances, walls, obstacles, and environmental changes to the remote team. Operators can inspect different directions as needed, without depending entirely on the vehicle’s current direction of travel.

This turns field inspection from a forward-facing video feed into a more complete site assessment. When a team needs to decide whether personnel should enter, which route to take, what risks to avoid, or whether the environment is still changing, 360° panoramic video provides stronger visual context for decision-making.

Facility and Critical Infrastructure Patrol: Preserving Spatial Context

In facilities, factories, energy sites, warehouses, hangars, communication stations, or other critical environments, patrol work is not only about recording fixed points. It is about understanding the continuous condition of the site.

An unmanned vehicle can follow planned routes to patrol perimeters, corridors, equipment zones, and restricted areas. With a single-direction camera, the patrol record may capture only what appears in front of the platform. However, an event may occur on the side, behind the vehicle, or shortly after the platform passes a certain point.

360° panoramic video helps patrol records preserve spatial context. When an abnormal situation is detected, or when an event needs to be reviewed later, operators can look back at the full surrounding scene before and after the event. This makes it easier to understand the location of the incident, movement direction, nearby personnel, and equipment changes.

For sites that require long-duration patrols, multi-point monitoring, or remote management across different areas, full-scene visual records can make remote operations more efficient and event review more reliable.

In mobile field operations, live video is often used by more than one operator. A management center, security team, engineering team, or decision-maker may all need to understand the on-site situation in real time.

Traditional video streaming often depends on the camera direction controlled by the front-line operator. Everyone watching the stream sees the same angle. When different roles need different information, a single viewpoint can limit collaboration.

360° panoramic video allows different users to enter the same visual scene and choose their own viewing direction. The operator can focus on vehicle movement, while remote personnel can inspect the side or rear environment. Engineers can look at equipment areas, while managers can quickly understand the overall field condition.

This changes remote collaboration from “watching the same camera angle” to sharing the same panoramic scene, where each team member can access the visual information most relevant to their task.

How Cupola360 Supports This Application

In mobile site awareness applications, Cupola360 provides the 360° panoramic vision layer, helping partner unmanned vehicles, robotic dogs, or autonomous mobile platforms capture broader surrounding video.

Compared with relying only on a front-facing camera, Cupola360 360° panoramic cameras allow remote operators to view the front, sides, and rear while the platform is moving. This helps reduce visual blind spots and preserves more complete field context for inspection, patrol, remote monitoring, and event review.

Cupola360 does not replace the unmanned vehicle or robotic platform. Instead, it provides the panoramic visual foundation that allows partners and system integrators to develop applications for inspection, reconnaissance, remote monitoring, event review, and site management based on richer visual context.

Related Resources

From Mobile Platforms to Reality Remote Management

The development of unmanned systems is creating new ways to manage sites that are high-risk, remote, or difficult for personnel to access continuously.

However, meaningful remote management is not only about sending video back to a control room. It is about enabling people to understand the site, assess events, and take action. This is the idea behind Cupola360’s Reality Remote Management: moving beyond monitoring to understand, manage, and act remotely.

When 360° panoramic vision is combined with unmanned systems, a mobile platform becomes more than a vehicle moving through a site. It becomes a visual node that continuously collects field context. This allows patrol, inspection, remote collaboration, and event review to be built on more complete on-site information.

Remote Situational Awareness: One Panoramic Scene for Multiple Teams

Application Highlights

Enables unmanned vehicles, robotic dogs, and mobile inspection platforms to capture 360° surrounding visibility
Reduces visual blind spots caused by front-facing cameras
Supports high-risk area inspection, facility patrol, critical infrastructure monitoring, and disaster site assessment
Allows remote operators to view the front, sides, and rear while the platform is moving
Preserves full field context for event review and abnormal situation assessment
Allows different teams to choose their own viewing direction within the same panoramic scene
Can be integrated with partner unmanned vehicles, robotic platforms, and system integration solutions

從遠端參訪到安全引導：資料中心的 360° 全景影像應用

Fri, 11 Sep 2026 09:13:37 GMT

結合全景相機、機器人、Cupola360⁺ Patrol 與即時 AI，讓參訪導覽與工程師進場流程更可控。

資料中心通常位於距離總部、客戶辦公室或合作夥伴據點較遠的園區。當外部訪客、合作夥伴或管理團隊需要了解機房環境、設備配置與營運狀態時，往往必須安排交通、接待、現場陪同與嚴格的出入流程。另一方面，進場施工或維護的工程師，也需要被引導至指定機櫃或作業區域，並確保全程遵守動線與存取規範。

在大型資料中心與高安全性機房場域中，這些流程不只耗費人力，也牽涉安全、效率與管理一致性。若每一次參訪或進場作業都高度依賴現場人員陪同，資料中心團隊就很難在不增加負擔的情況下，兼顧展示品質與安全控管。

透過自主移動機器人搭載 Cupola360 360° 全景相機，並結合 Cupola360⁺ Patrol 與即時 AI 分析，資料中心可以將遠端參訪、現場資訊呈現、工程師引導與安全監控整合在同一個操作流程中。附件中的方案架構也明確包含自主移動機器人、Cupola360 360° 相機、Patrol 平台與即時 AI 分析。

這項應用的重點，不只是把相機裝上機器人，而是讓資料中心從「遠端看見」進一步走向「遠端理解與管理」。遠端使用者不只能觀看現場，也能透過 360° 場景理解設備位置、區域資訊與營運狀態；進場人員則能在更少陪同人力下，維持清楚、可追蹤的安全管理流程。

資料中心參訪與工程作業，為什麼需要更可控？

資料中心是高度安全、管制嚴格的營運場域。無論是客戶參訪、合作夥伴展示、高階主管視察，或工程師進場維護，都不能像一般辦公空間一樣自由移動。

傳統參訪流程通常需要安排人員到場接待與全程陪同。這不只增加交通與人力成本，也讓參訪安排受到地點、時間、安全規範與現場營運條件限制。對海外客戶或跨區域團隊而言，安排一次完整現場參訪，往往需要更高的溝通成本與時程配合。

工程作業也面臨類似挑戰。資料中心內部機櫃密集，區域權限明確，工程師若不熟悉現場，很容易需要額外人員帶路。若進場人員偏離允許動線、誤入限制區域，或離開管理中心可掌握的範圍，就會增加安全管理與後續查核的壓力。

因此，資料中心需要的不只是遠端影像，而是一套能把現場畫面、場域資訊、導覽內容與安全控管結合起來的數位化流程。

遠端沉浸式參訪：不用到現場，也能理解資料中心

在遠端參訪情境中，訪客可在企業總部、遠端會議空間或展示中心，透過自主移動機器人搭載的 Cupola360 360° 全景相機瀏覽資料中心現場。

相較於傳統固定鏡頭或單一拍攝方向，360° 全景影像能讓使用者自由轉動視角，從不同方向理解機房空間、設備配置與周邊環境。這讓遠端參訪不只是觀看一段影片，而是更接近「身在現場」的沉浸式瀏覽體驗。附件也提到，訪客可透過機器人搭載的 Cupola360 鏡頭，以 360° 視角瀏覽資料中心，並自由轉動視角，不受傳統固定鏡頭限制。

透過 Cupola360⁺ Patrol，畫面中也能整合即時影像、設備狀態、區域名稱與導覽說明，讓參訪者在觀看場景的同時理解現場資訊。對資料中心業者而言，這有助於提升展示品質，讓客戶、合作夥伴或內部主管在不進入現場的情況下，也能快速掌握資料中心環境與營運亮點。

這種方式能減少跨地點交通、現場接待與陪同需求，也讓參訪安排更具彈性。尤其對海外客戶、跨區域團隊或高安全性機房而言，遠端參訪可以成為更有效率、更一致，也更容易管理的展示流程。

受控工程師引導：機器人帶路，AI 協助確認安全動線

方案架構：機器人、全景影像、平台與即時 AI

從遠端看見，走向遠端理解與管理

導入效益

Cupola360 如何支援這項應用

關鍵亮點

將 Cupola360 360° 全景相機部署於自主移動機器人上
支援大型資料中心與高安全性機房的遠端沉浸式參訪
讓使用者透過 360° 視角自由瀏覽資料中心現場
可在畫面中整合即時數據、設備狀態、區域名稱與導覽說明
透過機器人引導工程師前往指定機櫃或作業區域
即時 AI 可協助判斷人員是否離開機器人可視範圍或偏離允許動線
降低交通、接待與現場陪同成本
強化進場人員安全控管與後續查核依據
可延伸至多個園區、不同資料中心與高安全性場域

From Remote Tours to Safety Guidance: 360° Vision for Data Center Operations

Fri, 11 Sep 2026 07:29:23 GMT

Combining panoramic cameras, mobile robots, Cupola360⁺ Patrol, and real-time AI to make site visits and engineer guidance more controlled.

Data centers are often located in dedicated campuses away from corporate headquarters, customer offices, or partner locations. When external visitors, partners, or management teams need to understand the facility environment, equipment layout, or operational status, the process usually requires transportation, reception, on-site escorts, and strict access control.

Engineering work creates another challenge. Contractors or maintenance engineers entering a data center often need to be guided to specific racks or work areas while following strict movement and access rules.

In large data centers and high-security facilities, these workflows are not only resource-intensive. They also affect safety, efficiency, and operational consistency. If every visit or engineering task depends heavily on on-site escorts, it becomes difficult for data center teams to maintain both demonstration quality and access control without adding more operational burden.

By mounting Cupola360 360° panoramic cameras on autonomous mobile robots and combining them with Cupola360⁺ Patrol and real-time AI analysis, data centers can bring remote visits, site information display, engineer guidance, and safety monitoring into a single connected workflow.

The focus of this application is not simply placing a camera on a robot. It is about moving from remote viewing to remote understanding and management. Remote users can see the data center environment, understand equipment locations and site information, and interact with the scene through a 360° view. At the same time, on-site personnel can be guided with fewer escorting resources while maintaining a clearer and more traceable safety management process.

Why Data Center Visits and Engineering Work Need More Controlled Workflows

Data centers are high-security and tightly managed environments. Customer visits, partner demonstrations, executive reviews, and engineering maintenance cannot be handled like visits to a general office space.

Traditional visit workflows often require on-site reception and full-time escorts. This increases travel, coordination, and staffing costs. It also makes visits dependent on location, timing, security policies, and site operation conditions. For overseas customers or cross-region teams, arranging a full on-site visit can require significant planning and coordination.

Engineering workflows can be just as complex. Data center environments contain dense rack layouts, controlled areas, and strict access rules. Engineers who are unfamiliar with the site may need additional staff to guide them. If personnel move away from an approved route, enter a restricted area, or leave the visible range of the management center, it can create safety and audit challenges.

For this reason, data centers need more than remote video. They need a digital workflow that connects site visuals, contextual information, tour content, and safety control.

Immersive Remote Tours: Understanding the Data Center Without Being On-Site

In a remote tour scenario, visitors can explore the data center from a corporate headquarters, remote meeting room, or demonstration space through a Cupola360 360° panoramic camera mounted on an autonomous mobile robot.

Compared with fixed cameras or single-direction video, a 360° panoramic view allows users to freely look around and understand the facility space, equipment layout, and surrounding environment from different angles. This turns the remote visit from passive video watching into a more immersive site experience.

Cupola360⁺ Patrol can also display live video, equipment status, area names, and guided explanations within the same viewing interface. This allows visitors to understand site information while they are viewing the scene.

For data center operators, this can improve the quality and consistency of remote demonstrations. Customers, partners, and internal management teams can gain a better understanding of the facility without physically entering the site.

This approach helps reduce cross-location travel, on-site reception, and escorting requirements. It also makes visit scheduling more flexible, especially for overseas customers, distributed teams, and high-security facilities.

Controlled Engineer Guidance: Robot-Led Navigation with AI Safety Checks

Solution Architecture: Robot, Panoramic Vision, Platform, and Real-Time AI

From Remote Viewing to Remote Understanding and Management

Deployment Benefits

How Cupola360 Supports This Application

Related Resources

Another important application is controlled guidance for engineers entering the data center.

After an engineer enters the facility, the autonomous mobile robot can follow a predefined route and guide the person to a specific rack or work area. This helps reduce the risk of getting lost in a complex environment, entering a restricted zone by mistake, or requiring additional staff to escort the person throughout the entire process.

During movement, the Cupola360 360° panoramic camera provides a full surrounding view, allowing the management center to understand the on-site situation in real time. Real-time AI analysis can help determine whether the engineer remains within the robot’s visible range or stays on the approved route. If the person leaves the visible area or deviates from the permitted path, the system can trigger an alert.

This workflow strengthens personnel movement control and safety management without adding a large amount of escorting manpower. It also preserves video and event records that can support later review, audit, and process improvement.

This application brings together several key components:

Partner Autonomous Mobile Robot

Moves along designated routes and locations to support remote tours or on-site guidance.

Cupola360 360° Panoramic Camera

Provides a full surrounding view, reduces visual blind spots, and enables users to freely look around the scene.

Cupola360⁺ Patrol

Integrates live video, site information, equipment data, and tour content into one interface for remote viewing and site management.

Real-Time AI Analysis

Helps determine whether personnel remain within the robot’s visible range or approved movement route, and triggers alerts when needed.

With this architecture, data centers can convert visit and escort workflows that previously depended heavily on on-site personnel into a more standardized, traceable, and remotely manageable process.

Remote site tours should not be limited to streaming camera footage to a screen. For data centers, a valuable remote experience should allow users to see the site and understand what they are seeing.

This means the interface should show not only live video, but also area names, equipment status, guided explanations, and relevant real-time data. When a user sees a rack, aisle, or equipment area, they should also be able to understand its location, function, and operational context.

Similarly, engineer guidance is not only about having a robot lead the way. It is about allowing the management center to understand whether personnel are following the correct route, staying within approved areas, and remaining visible during the process.

This is where 360° panoramic vision becomes valuable. Its role is not only to provide a wider view, but to bring physical space, real-time information, and safety workflows into the same operational context.

Reduced Visit Costs

Reduces cross-location travel, on-site reception, and escorting requirements, making customer visits, partner demonstrations, and internal reviews more flexible.

Improved Demonstration Quality

Combines 360° free-viewing with real-time data and guided information, allowing visitors to see the environment and understand operational context at the same time.

Stronger Safety Control

Uses robot guidance and AI-assisted monitoring to reduce the risk of personnel leaving the approved route, entering restricted zones, or moving out of the visible range.

Repeatable Workflow

Standardizes remote visits and engineer guidance, allowing the workflow to be extended to multiple campuses, different data centers, or other high-security facilities.

In this data center application, Cupola360 provides 360° panoramic cameras and Cupola360⁺ Patrol, enabling a partner autonomous mobile robot to capture full surrounding video while moving through the site and bring panoramic views, site information, and guided content into a remote operating interface.

With 360° panoramic vision, remote users can not only view the data center environment, but also freely look around and understand the spatial relationship between racks, aisles, and equipment areas. When combined with real-time AI analysis, the system can also help determine whether on-site personnel have left the robot’s visible range or deviated from an approved route.

Cupola360’s value is not to replace the robot or existing management systems. It provides the 360° visual foundation and remote viewing experience that allows partners and data center operation teams to build workflows for remote tours, safety guidance, and site management.

Key Highlights

Deploys Cupola360 panoramic cameras on autonomous mobile robots
Supports immersive remote tours for large data centers and high-security facilities
Allows users to freely explore the data center through a 360° view
Can display live data, equipment status, area names, and guided explanations
Guides engineers to designated racks or work areas through a mobile robot
Uses real-time AI to help determine whether personnel leave the robot’s visible range or deviate from approved routes
Reduces travel, reception, and on-site escorting requirements
Strengthens personnel safety control and supports later review
Can be extended to multiple campuses, data centers, and high-security environments

IP66 vs IP67｜戶外全景相機防水防塵等級怎麼選

Tue, 08 Sep 2026 07:25:29 GMT

防護等級不是數字越大就一定越適合，關鍵在於設備實際部署時會面對什麼環境。

在戶外相機、工業設備與監控產品規格中，IP 防護等級是客戶經常關注的項目之一。許多人看到 IP66、IP67 或 IP68 時，會直覺認為數字越大代表規格越高，也代表產品越耐用。

但在實際部署中，IP 防護等級並不是單純的數字比較。不同 IP 等級代表不同測試情境，也對應不同環境風險。對戶外全景相機而言，真正重要的不是追求更大的數字，而是選擇最符合實際使用情境的防護設計。

這也是為什麼 Cupola360 選擇以 IP66 作為戶外產品的重要防護標準。

IP 防護等級代表什麼？

IP 等級是用來描述設備外殼對固體異物與水侵入的防護能力。一般常見格式為 IPXX，第一個數字代表防塵等級，第二個數字代表防水等級。

以 IP66 與 IP67 為例：

第一個數字 6 代表防塵等級，表示設備具有高度防塵能力。
第二個數字則代表防水測試情境不同。

也就是說，IP66 與 IP67 在防塵等級上相同，但在防水測試上，兩者重點並不一樣。

IP66 與 IP67 的差異

IP66 著重於設備抵抗強力水柱的能力。這類測試主要驗證產品在豪雨、強風夾帶雨水、高壓水柱沖洗或戶外清潔等情境下，外殼是否仍能維持良好的密封性能。

IP67 則著重於短時間浸水測試。一般來說，IP67 驗證的是設備可在約 1 公尺深的靜止水中浸泡約 30 分鐘，適合可能發生短時間浸水、掉入水中或暫時被水覆蓋的應用情境。

因此，IP66 與 IP67 並不是簡單的「誰比較高級」。它們測試的是不同風險：

值得注意的是，IP67 著重浸水測試，並不等同於自動通過 IP66 的強力水柱測試。換句話說，具備 IP67 認證，不代表產品一定符合 IP66 的使用情境。

對多數戶外相機來說，最常見的環境挑戰通常不是「整台設備被泡在水裡」，而是：

長時間暴露在雨水中
遇到豪雨或颱風天強風帶來的雨水衝擊
空氣中的粉塵、砂粒或污染物
現場清潔時的水柱沖洗
日夜溫差與戶外濕氣

在這些情境中，設備需要抵抗的是持續噴濺、強力水柱、雨水衝擊與粉塵侵入。這些需求更接近 IP66 的測試重點。

如果一台戶外相機主要安裝在建築外牆、路口、工廠外圍、停車場、園區、觀光景點或公共空間，正常情況下並不會長時間浸泡在水中。這時候，能否抵抗強風豪雨、水柱清潔與戶外粉塵，往往比短時間浸水測試更貼近實際使用情境。

為什麼 Cupola360 選擇 IP66？

Cupola360 的產品設計重點，是支援真實戶外與工業場域中的長時間穩定運作。這些場域可能包含工廠、倉儲、停車場、校園、公共空間、觀光景點、交通場域與戶外監控點。

在這些應用中，相機需要面對的通常是豪雨、粉塵、強風、環境噴濺與設備清潔，而不是水下作業或長時間浸泡。因此，IP66 更符合 Cupola360 戶外全景相機的主要部署需求。

選擇 IP66，代表我們更重視產品在常見戶外嚴苛環境中的實用防護能力。對大多數戶外監控與遠端場域管理應用而言，IP66 已能提供可靠的防水防塵基礎，協助設備在不同氣候與現場條件下維持穩定運作。

為什麼戶外相機不一定需要 IP67？

IP67 什麼時候才會比較重要？

如果應用場景涉及短時間浸水、積水風險、水下作業，或設備可能被水覆蓋，那麼 IP67、IP68 或額外防水殼設計就可能更重要。

例如：

設備可能安裝在會積水的位置
相機可能被短時間淹沒
應用接近水池、水槽、河道或水下環境
專案有明確的浸水測試要求

這類情境就不能只用一般戶外防雨需求來判斷，而應依據實際環境條件，選擇更合適的防護方案。對特殊環境而言，也可以搭配額外防水殼或客製化安裝設計，進一步滿足專案需求。

選擇 IP 等級時，應該看什麼？

在評估戶外相機防護等級時，建議不要只比較 IP 數字，而是回到實際部署環境：

設備是暴露在雨中，還是可能被水浸泡？
現場是否會使用水柱沖洗設備？
是否有粉塵、砂粒、油霧或污染物？
安裝位置是否容易積水？
是否需要長時間戶外運作？
是否有特殊產業規範或客戶驗收標準？

只要回到這些問題，就會更容易理解：IP66 與 IP67 並不是單純高低差異，而是不同使用情境下的防護選擇。

結論

對戶外全景相機而言，防護設計的重點不是追求更大的 IP 數字，而是選擇最符合實際應用需求的防護能力。

IP66 著重強力水柱、豪雨與戶外噴濺，適合大多數戶外監控、工業場域與遠端場域管理應用。IP67 則著重短時間浸水，適合可能發生浸泡或落水風險的設備。

Cupola360 選擇 IP66，是因為我們的產品主要面對的是真實戶外部署環境：雨水、粉塵、強風、清潔沖洗與長時間運作需求。真正重要的不是數字看起來更大，而是防護標準是否真正符合設備會面對的現場挑戰。

IP66 vs. IP67 | Protection Ratings for Outdoor Panoramic Cameras

Tue, 08 Sep 2026 07:10:47 GMT

A higher IP number does not always mean a better fit. The right protection rating depends on the environment the device is designed to face.

For outdoor cameras, industrial equipment, and monitoring devices, IP protection rating is one of the most common specifications customers look at. When people see IP66, IP67, or IP68, it is easy to assume that the larger number always means a higher level of protection.

In real-world deployment, however, IP ratings are not simply a ranking system. Different IP ratings represent different test conditions, and those test conditions are designed for different environmental risks. For outdoor panoramic cameras, the key question is not which number looks higher, but which protection standard best matches the actual application.

This is why Cupola360 uses IP66 as an important protection standard for its outdoor products.

What Does an IP Rating Mean?

An IP rating describes how well a device enclosure protects against solid particles and water ingress. The format is usually written as IPXX, where the first digit refers to protection against dust and solid objects, and the second digit refers to protection against water.

For IP66 and IP67:

The first digit 6 indicates a high level of dust protection.
The second digit indicates different water protection test conditions.

In other words, IP66 and IP67 share the same dust protection level, but they are tested against different types of water exposure.

The Difference Between IP66 and IP67

IP66 focuses on protection against powerful water jets. This test condition is relevant for heavy rain, wind-driven water, high-pressure cleaning, outdoor splashing, and equipment washdown scenarios.

IP67 focuses on temporary immersion. It is generally used to verify whether a device can withstand being submerged in still water at a depth of about 1 meter for approximately 30 minutes. This is more relevant for applications where a device may be briefly immersed, dropped into water, or temporarily covered by water.

This means IP66 and IP67 are not simply “lower” and “higher” versions of the same requirement. They address different risks:

It is also important to note that IP67 focuses on immersion testing and does not automatically mean the product has passed the powerful water jet test required for IP66. A device with IP67 certification is not necessarily suitable for all IP66 use cases.

For most outdoor cameras, the most common environmental challenge is not full submersion. Instead, the device is more likely to face:

Long-term exposure to rain
Heavy rain or wind-driven water during storms
Dust, sand, or airborne particles
Water jets during site cleaning
Outdoor humidity and temperature variation

In these scenarios, the device must withstand rain, splashing, powerful water jets, dust ingress, and long-term outdoor exposure. These conditions are more closely aligned with IP66 testing.

If an outdoor camera is installed on a building exterior, road intersection, factory perimeter, parking area, campus, tourist site, or public space, it is typically not designed to remain underwater. In this type of deployment, resistance to heavy rain, water jets, outdoor dust, and cleaning conditions is often more relevant than temporary immersion.

Why Cupola360 Chooses IP66

Cupola360 products are designed to support long-term operation in real outdoor and industrial environments. These may include factories, warehouses, parking areas, campuses, public spaces, tourist sites, transportation locations, and outdoor monitoring points.

In these applications, cameras are more likely to encounter heavy rain, dust, strong wind, splashing, and equipment cleaning, rather than underwater operation or temporary submersion. For this reason, IP66 is better aligned with the primary deployment scenarios of Cupola360 outdoor panoramic cameras.

By choosing IP66, Cupola360 focuses on practical protection for common outdoor harsh environments. For most outdoor monitoring and remote site management applications, IP66 provides a reliable weatherproof and dustproof foundation, helping devices maintain stable operation under different site and climate conditions.

Why Outdoor Cameras Do Not Always Need IP67

When Does IP67 Matter More?

If an application involves temporary immersion, standing water, underwater operation, or a risk that the device may be covered by water, then IP67, IP68, or an additional waterproof enclosure may be more important.

Examples include:

Installation locations with potential flooding or standing water
Cameras that may be temporarily submerged
Applications near pools, tanks, rivers, or underwater environments
Projects with specific immersion test requirements

In these cases, the protection design should not be evaluated as a general outdoor rainproof requirement. It should be based on the actual environmental risk and project specifications.

For special environments, additional waterproof housings or customized installation designs can also be considered to meet project requirements.

What to Consider When Choosing an IP Rating

When evaluating the protection rating for an outdoor camera, it is better not to compare IP numbers alone. Instead, start from the actual deployment environment:

Will the device be exposed to rain, or could it be submerged?
Will the site use water jets for equipment cleaning?
Is there dust, sand, oil mist, or other airborne contamination?
Is the installation point likely to accumulate water?
Does the device need to operate outdoors for long periods?
Are there specific industry standards or customer acceptance requirements?

These questions make it easier to see why IP66 and IP67 are not simply different levels of the same requirement. They are different protection choices for different use conditions.

Conclusion

For outdoor panoramic cameras, protection design is not about chasing the largest IP number. It is about choosing the protection capability that best matches the actual application.

IP66 focuses on powerful water jets, heavy rain, and outdoor splashing, making it suitable for most outdoor monitoring, industrial site, and remote site management applications. IP67 focuses on temporary immersion, making it more suitable for devices that may face submersion or water coverage.

Cupola360 chooses IP66 because its products are primarily designed for real outdoor deployment conditions: rain, dust, wind, washdown, and long-term operation.

The real question is not which number looks higher. It is whether the protection standard matches the challenges the device will actually face in the field.

Related Resources

從監控畫面到營運洞察：百貨公司如何用全景相機支援 AI 影像分析

Tue, 01 Sep 2026 07:12:47 GMT

讓人流、熱區與停車影像成為百貨營運決策的參考依據。

在一項大型百貨公司 POC 中，Cupola360 提供 360° 全景相機作為現場影像來源，搭配第三方 AI 影像分析平台，協助驗證百貨場域如何將影像資料應用於人流分析、場域感知、安防自動化與營運規劃。

對百貨公司而言，挑戰通常不是「有沒有攝影機」。主出入口、戶外看板、店內活動展區、停車場與公共空間，往往都已有不同影像來源。真正的問題是，這些影像能不能從單純監看與回放，進一步轉化成營運團隊可以使用的決策資訊。

挑戰：百貨公司需要的不只是監控，而是跨場域營運理解

大型百貨公司的現場管理，往往牽涉多個彼此關聯的區域。入口人流會影響樓層導引與人力配置；戶外看板的停留狀況會影響廣告排播與媒體效益；店內展區的人潮熱點會反映活動吸引力；停車場車流則能揭示不同客群的到訪節奏。

如果這些資訊分散在不同攝影機、不同報表或不同管理系統中，營運團隊就很難快速看出關聯。現場看起來很熱鬧，不代表每個展示區都有效；停車場尖峰，也不一定與主出入口人流尖峰同步；看板前有人經過，也不代表真正產生停留與注意力。

因此，百貨公司需要的不只是更多影像，而是一套能把影像、事件、人流、停留、熱區與車流資料連接起來的視覺分析架構。當現場資料能被整理成可判讀的營運訊號，影像才不只是記錄工具，而能進一步成為百貨營運與行銷優化的決策基礎。

解決方案：跨區域 AI 影像分析 POC 驗證

本次 POC 針對大型百貨公司的多個典型場域進行 AI 影像分析驗證，包括主出入口、戶外廣告看板區、店內活動展區與地下停車場。對應應用包含人員滯留偵測、人群聚集告警、環境異常監控、行人流量統計、車牌辨識、車輛計數、自動化報表產出、客層特徵分析與熱區分布圖。

透過 AI 影像分析，原本分散的現場畫面被轉化為可量化、可回顧、可延伸應用的營運資料。百貨團隊不只可以查看現場，也能進一步理解人流節奏、顧客停留、區域熱度、車流高峰與安全事件。

在 POC 驗證中，9 項關鍵任務完成測試並通過驗證，展現系統在複雜零售環境中的適應性與穩定性。

主出入口：建立人流與客層基準

戶外看板區：從曝光位置走向停留分析

Key Highlights

Related Resources

主出入口是百貨公司最重要的營運數據入口之一。透過 AI 影像分析，系統可即時統計人員進出，並產出進入、離開與總量等統計資料。相關報表可依小時、天、週、月等時間維度整理，並匯出 Excel 供營運分析使用。

這類資料可以協助百貨營運團隊建立人流基準，理解不同時段、不同日期與活動期間的人潮變化。當人流資料與客層輪廓結合後，也能進一步支援活動規劃、樓層導引、人力配置與行銷策略。

在本次 POC 中，主出入口同時驗證了行人計數、年齡／性別分析與客層分析應用。

正式對外溝通時，建議將這類能力描述為匿名化客層輪廓分析或人流營運分析，避免過度強調個人辨識。對百貨業者而言，真正有價值的不是追蹤單一個人，而是理解整體客流、族群輪廓與時段變化。

戶外看板的價值，不只是有多少人經過，更重要的是顧客是否真的停留、觀看，甚至在特定位置形成聚集。

在本次 POC 中，戶外廣告看板區設定人員滯留偵測。當人員在偵測範圍內停留超過設定秒數時，系統即可產生告警紀錄。

這類資料能幫助百貨與零售媒體團隊理解看板周邊的停留強度，進一步評估廣告位置、視覺設計與排播時段。過去看板成效可能只能透過人流估算或曝光位置推測，但透過影像分析，團隊可以更接近真實現場行為，知道哪些時段、哪些位置更容易產生停留。

簡報中也提出，可依滯留高峰規劃動態廣告排播，例如在人潮停留較密集的時段播放品牌形象內容，或在停車族入場高峰時段播放樓層導引與促銷資訊。

這讓戶外看板不再只是靜態媒體，而是能與現場人流行為連動，成為更具資料基礎的零售媒體入口。

針對大型百貨公司場域進行 POC 驗證
覆蓋主出入口、戶外廣告看板、店內活動展區與停車場等多類場景
驗證人流統計、客層輪廓、滯留偵測、人群聚集、熱區分析、車牌/車型辨識與車輛計數等應用
9 項關鍵任務完成測試並通過驗證
支援即時告警，告警觸發延遲小於 1 秒
將影像資料轉化為人流、停留、熱區、車流與安全事件洞察
可延伸至自動化報表、POS 轉換率分析、動態廣告排播與統一 AI 影像營運中心
適用於百貨公司、購物中心、大型零售、展區活動與多分店場域管理

店內活動展區：用熱區分析看見真正吸引顧客的位置

店內活動展區是本次 POC 中最具零售營運價值的場景之一。

透過人群聚集偵測，系統可在指定區域內人數超過設定門檻時，自動發出告警並記錄事件。

而熱區分析則可依每小時或彙整期間產出熱區圖，統計影像中人員或物件的分布情況，並可匯出 Excel 供後續分析。

對百貨營運與活動企劃團隊而言，熱區圖的價值在於，它可以把「感覺很熱鬧」變成更具體的資料。團隊可以看到哪些展示攤位停留最久、哪些區域只是通行動線、哪些位置需要調整陳列或導引。

在 POC 分析中，特定展示區被判定為最強熱點，顧客挑選商品與拍照打卡行為使停留時間明顯高於其他區域；同時也觀察到特定時段是活動、快閃或限定商品開賣的高互動窗口。

這類資訊可支援品牌快閃、展區配置、商品陳列、動線規劃與活動排程，讓百貨空間規劃更接近真實顧客行為。

停車場：把車流與顧客行為放進同一張營運地圖

停車場經常是大型百貨場域中被低估的數據來源。事實上，車流進出、車型組成、尖峰時段與停留狀況，都能反映不同客群的到訪模式。

在本次 POC 中，停車場區域驗證了人員滯留告警、環境異常監控、車牌與車型辨識、車輛計數，以及火煙偵測等功能。

車輛計數可依不同時間維度進行統計，並區分多種車型，包括腳踏車、大貨車、汽車、機車與貨卡車。

這能協助營運團隊理解停車場壓力、尖峰進場時段與不同交通方式帶來的客流差異。POC 分析也指出，停車尖峰與主出入口人流尖峰可能不完全重疊，顯示開車族與大眾運輸族可能有不同的到訪節奏與消費行為。

對百貨公司而言，這類跨場域洞察特別有價值。它不只可以協助停車場管理，也能支援分眾行銷、入口導引、樓層活動安排與促銷時段規劃。

從影像監控走向百貨營運洞察

本次 POC 的價值，不只在於完成單一 AI 功能驗證，而是展現影像分析如何把不同百貨場域串接起來。

主出入口提供人流節奏。

戶外看板提供停留與曝光線索。

店內活動展區提供熱區與互動強度。

停車場提供車流、車型與安全事件資料。

當這些資料被放在同一個營運脈絡中，百貨團隊就能更清楚理解顧客如何進入、移動、停留、互動與離開。影像不再只是事後回看的工具，而是能支援營運決策、行銷優化與場域管理的資料來源。

簡報後續導入規劃也提到，可從基礎建設與安防出發，逐步建立自動化每日人流報表、整合停車場管理系統，再延伸到 POS 轉換率、動態廣告排播、熱區定期分析，最後擴展至更多分店與統一 AI 影像營運中心。

這也說明了百貨場域導入 AI 影像分析時，不一定要一次到位，而是可以從關鍵場域驗證開始，逐步走向更完整的智慧營運架構。

系統驗證與部署彈性

在技術驗證方面，本次 POC 達成即時處理能力，告警觸發延遲小於 1 秒，並支援多路影像同步分析，不影響系統穩定性。

同時，實際導入時仍需針對環境進行調校。例如戶外強光、停車場低光源、複雜背景與誤報率控制，都是零售場域 AI 影像分析能否成功落地的重要關鍵。

本次驗證採用短影像模擬串流進行功能驗證，但技術架構與未來實體 camera 對接邏輯一致，因此具備高度可遷移性與擴展性。

Cupola360 如何支援這項應用

在此百貨公司 POC 中，Cupola360 提供 360° 全景相機作為現場影像來源，讓第三方 AI 影像分析平台能取得更完整的場域視角，進一步進行人流、停留、熱區、車流與事件分析。

相較於只依賴固定視角攝影機，360° 全景相機能從單一位置提供更寬廣的現場畫面，協助百貨場域在出入口、活動展區、戶外看板與停車場等不同區域，取得更完整的影像基礎。

在這類應用中，Cupola360 的價值不是取代既有監控系統或 AI 平台，而是提供更完整的全景影像來源，讓系統整合商、AI 影像分析平台與百貨營運團隊，能在更完整的視覺資料上建立人流分析、熱區觀察、停車行為分析與場域管理流程。

From Video Monitoring to Store Insights: How Panoramic Cameras Support Department Store AI Analytics

Tue, 01 Sep 2026 06:51:52 GMT

Turning department store video into customer flow, heatmap, parking, and operation insights.

In a proof-of-concept for a large department store environment, Cupola360 provided 360° panoramic cameras as the site video source, working with a third-party AI video analytics platform to evaluate how visual data could support customer flow analysis, site awareness, security automation, and operation planning.

For department stores, the challenge is rarely a lack of cameras. Entrances, signage areas, exhibition zones, parking facilities, and public spaces are often already covered by different video sources. The real question is whether those video feeds can move beyond monitoring and playback to become useful operational insight.

The Challenge: Department Stores Need More Than Monitoring

Large department stores are complex operating environments. Visitor flow at the entrance affects wayfinding and staffing. Dwell activity near outdoor signage influences media value and advertising schedules. In-store heatmaps reveal which displays attract attention. Parking data helps explain arrival behavior and visitor patterns.

When these signals are scattered across separate cameras, reports, or management systems, operation teams may struggle to understand how they connect. A busy-looking area does not always mean every display is effective. Parking peaks may not align with entrance traffic peaks. A high-traffic signage area does not necessarily mean visitors are stopping or engaging.

This is why department stores need more than additional video feeds. They need a connected visual analytics architecture that can bring together video, events, customer flow, dwell behavior, heatmaps, and parking data.

When visual data becomes structured and readable, video is no longer only a record of what happened. It becomes a decision-making layer for department store operations, marketing optimization, and site management.

The Solution: Multi-Zone AI Video Analytics Validation

This POC validated AI video analytics across multiple department store zones, including main entrances, outdoor signage areas, in-store exhibition zones, and underground parking facilities. The applications included dwell detection, crowd gathering alerts, environmental anomaly monitoring, pedestrian traffic counting, license plate recognition, vehicle counting, automated reporting, audience profile analysis, and heatmap visualization.

Instead of treating video as isolated footage, the system converted visual inputs into measurable, reviewable, and extensible operational data. This helped the retail team understand traffic rhythm, dwell behavior, area engagement, parking flow, and safety-related events across different parts of the site.

During the POC, nine key tasks were completed and passed validation, demonstrating system adaptability and stability in a complex retail environment.

Main Entrance: Building a Baseline for Customer Flow

Outdoor Signage: Turning Exposure into Dwell-Time Insight

Key Highlights

Related Resources

The main entrance is one of the most important operational data points in a department store. Through AI video analytics, the system can count people entering and exiting in real time, generate total traffic numbers, and support reporting by hour, day, week, or month. The data can also be exported to Excel for operational analysis.

This helps department store teams establish a customer flow baseline and understand how visitor volume changes across different time periods, dates, campaigns, or events. When traffic data is combined with audience profile analysis, it can further support staffing, event planning, wayfinding, floor guidance, and marketing decisions.

In this POC, the main entrance area validated pedestrian counting, age and gender analysis, and audience profile applications.

For public-facing communication, this type of capability should be described as anonymous audience profile analysis or customer flow analytics, rather than emphasizing personal identification. For department store operators, the real value is not tracking a specific person. It is understanding overall visitor flow, audience composition, and time-based behavior patterns..

Outdoor signage is not only about how many people pass by. For retail media and brand communication, a more useful question is whether visitors actually stop, stay, and engage with the display area.

In the POC, the outdoor signage zone was configured for dwell detection. When a person stayed within the detection area beyond a defined threshold, the system generated an alert record.

This type of data helps department store and retail media teams evaluate signage performance, media placement, visual design, and content scheduling. Instead of relying only on estimated exposure or location-based assumptions, operators can better understand when and where people are more likely to stop.

The report also suggested that dwell-time peaks could be used to support dynamic advertising schedules, such as displaying high-impact brand content during high-dwell periods or using specific time windows to promote floor guidance and retail offers.

With AI video analytics, outdoor signage can become more than a static media asset. It can become a data-informed touchpoint connected to real customer behavior.

Conducted a POC in a large department store environment
Covered main entrances, outdoor signage, in-store exhibition zones, and parking facilities
Validated people counting, anonymous audience profile analysis, dwell detection, crowd gathering alerts, heatmap analysis, license plate and vehicle recognition, and vehicle counting
Completed and passed nine key validation tasks
Supported real-time alerts with less than one-second trigger latency
Turned visual data into insights on customer flow, dwell behavior, heatmaps, vehicle flow, and site alerts
Can be extended to automated reporting, POS conversion analysis, dynamic advertising schedules, and unified AI video operations
Suitable for department stores, shopping malls, large retail sites, exhibition zones, and multi-location retail operations

In-Store Exhibition Zone: Understanding What Actually Attracts Visitors

The in-store exhibition zone was one of the most valuable retail operation scenarios in this POC.

Crowd gathering detection can automatically trigger alerts when the number of people in a defined area exceeds a preset threshold.

Heatmap analysis can also generate hourly or summarized heatmap views, showing where people or objects are distributed within the video scene, with Excel export available for further analysis.

For department store operation and event planning teams, heatmaps help turn subjective impressions into measurable insight. Instead of simply saying an area “felt crowded,” operators can identify which display zones attracted the most attention, which areas were mainly used as walkways, and which layouts may need adjustment.

The POC analysis showed that specific display areas generated stronger dwell activity, while certain time periods became high-engagement windows for product launches, limited-time offers, or pop-up activities.

This type of insight can support pop-up planning, product display optimization, campaign scheduling, traffic guidance, and retail space design.

Parking Area: Connecting Vehicle Flow with Customer Behavior

Parking facilities are often an overlooked source of department store intelligence. In reality, parking traffic, vehicle types, peak arrival periods, and dwell behavior can reveal important differences in visitor patterns.

In this POC, the parking area validated multiple applications, including dwell alerts, environmental anomaly monitoring, license plate and vehicle model recognition, vehicle counting, and fire or smoke detection.

Vehicle counting can support reporting by different time ranges and classify multiple vehicle types, including bicycles, trucks, cars, motorcycles, and pickups.

This helps operation teams understand parking pressure, peak arrival periods, and differences between visitor groups. The POC analysis also found that parking peak hours and entrance traffic peak hours may not fully overlap, suggesting that drivers and public-transport visitors may follow different arrival and shopping patterns.

For department store operators, this type of cross-zone insight is especially valuable. It can support parking operations, segmented marketing, entrance-specific wayfinding, floor activity planning, and time-based promotion strategies.

From Video Monitoring to Department Store Site Intelligence

The value of this POC was not only in validating individual AI functions. Its greater value was showing how different department store zones can be connected through visual intelligence.

The entrance provides traffic rhythm.

Outdoor signage provides dwell and exposure signals.

In-store exhibition zones provide engagement and heatmap data.

Parking areas provide vehicle flow, visitor pattern, and safety event information.

When these signals are viewed together, department store teams can better understand how visitors enter, move, stay, interact, and leave. Video becomes more than a playback tool. It becomes a data source for operations, marketing optimization, and site management.

The proposed next steps in the POC included connecting live cameras, building automated daily traffic reports, integrating parking management systems, connecting POS data for conversion analysis, supporting dynamic advertising schedules, conducting regular heatmap analysis, expanding to additional locations, and building a unified AI video operations center.

This also shows that AI video analytics does not need to be deployed all at once. Department store operators can start from key zones, validate priority use cases, and gradually expand into a more complete intelligent operation workflow.

System Validation and Deployment Flexibility

From a technical validation perspective, the POC achieved real-time processing, with alert-trigger latency of less than one second. It also supported multi-channel video analysis without affecting system stability.

The report also noted that real-world deployment requires site-specific tuning. Outdoor lighting, low-light parking conditions, complex backgrounds, and false-positive reduction all need to be considered when deploying AI video analytics in department store environments.

The validation used short video clips to simulate streaming, while the technical architecture remained consistent with future physical camera integration. This means the validated workflow can be migrated and scaled when connected to real camera sources.

How Cupola360 Supports This Application

In this department store POC, Cupola360 provided 360° panoramic cameras as the site video source, enabling a third-party AI video analytics platform to access broader visual coverage for customer flow, dwell, heatmap, vehicle flow, and event analysis.

Compared with fixed-view cameras alone, 360° panoramic cameras can capture a wider site context from a single installation point. This helps department store environments collect more complete visual input across entrances, exhibition zones, outdoor signage areas, parking facilities, and other operational spaces.

In this type of application, Cupola360 does not replace the existing VMS, AI analytics platform, or management system. Instead, it provides a panoramic video foundation that allows system integrators, AI video analytics platforms, and department store operation teams to build stronger workflows for customer flow analytics, heatmap observation, parking behavior analysis, and site awareness.

DP Smart × Cupola360：以 360° 即時影像打開智慧城市與公共場域的全景視野

Fri, 28 Aug 2026 09:07:59 GMT

從觀光景點、公共空間、交通節點到展覽活動與社區賽事，360° 即時串流讓多元場域形成可觀看、可分享、可延伸管理的即時視覺網路。

DP Smart 達標智源是 Cupola360 在台灣的重要合作夥伴之一，長期投入 360° 即時影像應用部署，應用範圍涵蓋觀光景點、公共空間、交通節點、活動展場、運動賽事與各類需要遠端可視性的開放場域。

透過 Rogy360 / 360 See Taiwan 等即時影像應用，DP Smart 展示了 360° 即時串流不只是單一景點的 webcam 直播，而是能夠被規模化部署，形成橫跨不同場域的公開即時視覺網路。

從國家公園、風景區、老街、廟宇、城市地標，到交通路口、展覽現場、電競活動、公共場館與社區賽事，360° 全景直播讓人們可以不受地點限制，直接看見現場當下的環境狀況。對城市、場域經營者與系統整合夥伴而言，這也代表一種更具彈性的智慧城市視覺基礎。

挑戰：公共場域需要更即時、更直覺的現場資訊

公共場域具有高度動態特性。天氣會變、人潮會聚集、交通狀況會改變，活動現場也會隨時間呈現不同狀態。

對旅客、市民、活動參與者或遠端觀看者來說，靜態資訊往往不夠。一張照片可以呈現某個地點的樣貌，卻無法說明現場現在的天氣、人流、能見度、交通狀況或活動氛圍。文字公告能提供基本資訊，但很難完整傳達現場的空間感與即時狀態。

對場域經營者與公共空間管理者而言，挑戰也不只是對外資訊揭露。許多公共場域分散在不同地點，例如觀光景點、交通節點、道路路口、展覽場館、公園、遊客中心與活動空間。若完全仰賴人工巡查或單一固定視角，不僅成本較高，也容易造成資訊零散與反應延遲。

360° 即時串流提供了一種更直覺的方式，讓這些場域能被即時看見。

從景點直播到公共場域即時視覺網路

DP Smart 的 360° 即時串流模式，不只適用於觀光景點。觀光場域是重要應用之一，但更大的價值在於，這套模式可以被複製到不同類型的公開環境，形成可持續營運的即時影像網路。

單一 360° 即時影像可以幫助使用者理解一個地點；多個 360° 即時影像則能讓使用者、管理者與合作夥伴持續掌握多個地點的現場狀態。

這讓 live streaming 不再只是媒體內容，而是成為公共場域的一層即時視覺基礎。觀看者可以打開直播、環顧現場，並在不親自抵達的情況下理解當下環境。管理單位也可以將同一類影像來源作為遠端掌握、事件回顧或未來 AI 應用的基礎。

在這樣的架構下，360° 即時串流可以成為智慧城市與公共空間應用中的一種可擴充視覺基礎設施。

可應用的場域分類

為什麼 YouTube Live 適合公開觀看

從公開觀看延伸到管理端整合

Cupola360 的角色：全景視覺基礎

打開智慧城市的即時全景視野

關鍵亮點

展示 360° 即時串流如何橫跨不同公共場域進行規模化部署
應用範圍從觀光景點延伸至公共空間、交通節點、活動展場、運動賽事、公園與社區活動
透過 YouTube Live 作為公開觀看與內容分發渠道
讓旅客、市民與線上觀眾能輕鬆開啟 360° 即時影像
支援觀光推廣、公共溝通、活動曝光與遠端場域掌握
以 Cupola360 全景相機系列作為視覺基礎
可依專案需求延伸至 Cupola360 ⁺ Patrol，支援管理端監控與系統整合
支援智慧城市與公共場域的即時全景視覺應用模式

備註 : 體驗更多不同場域的 360° 全景直播

零售門市案例 : 360° 全景視覺支援收銀作業監控

Fri, 28 Aug 2026 03:49:16 GMT

透過全景門市視角、AI 輔助偵測與收銀區影像整合，協助零售團隊更有效掌握收銀作業、事件回顧與門市營運狀況。

在零售門市中，收銀區是最關鍵、也最容易產生爭議的作業區域之一。每一次交易、開啟錢櫃、現金處理、交班與異常金額差異，都可能需要清楚的影像脈絡來協助釐清狀況。

傳統監視器通常會針對收銀機、櫃台或門口設置固定角度畫面，能提供局部細節；但若要理解整個事件，單一畫面往往不夠。管理者可能還需要知道當時門市內有多少人、是否有其他人靠近收銀區、員工與顧客的互動情況，以及事件發生前後的完整現場脈絡。

在這個零售應用案例中，Cupola360 360° 全景視覺與 AI 輔助偵測、傳統收銀區攝影機搭配使用，協助門市團隊更有效監控收銀作業流程、回顧交易爭議，並提升遠端門市管理效率。

挑戰：收銀作業需要細節，也需要完整現場脈絡

收銀區的管理挑戰，不只是確認某一筆交易有沒有完成。對零售團隊而言，真正需要掌握的是整個收銀流程是否符合規範，以及當異常發生時，是否能快速還原當下狀況。

例如，當現金對帳出現差異時，管理者可能需要回看當時錢櫃開啟的時間、收銀人員是否在場、是否有其他員工或顧客靠近櫃台，以及交易前後是否有異常行為。

如果影像只來自單一固定角度，可能看得到收銀機畫面，卻看不到周邊人員動線；看得到櫃台正面，卻無法理解店內其他位置發生什麼事。這會讓事件回顧變得耗時，也增加判斷上的不確定性。

因此，零售門市需要的不只是更多鏡頭，而是能把局部細節與全店脈絡結合起來的視覺架構。

解決方案：全景門市視角結合收銀區畫面

在此應用中，360° 全景相機提供門市整體視野，協助管理者從單一畫面掌握店內人員位置、動線與收銀區周邊狀況。傳統固定攝影機則可用於收銀機或櫃台細節畫面，補足局部特寫需求。

透過這樣的搭配，操作人員可以先從 360° 全景畫面理解整體門市狀況，再進一步點選或查看收銀區畫面，確認交易細節與人員互動。

這種視覺架構特別適合連鎖零售、便利商店、生活用品店、超市、餐飲收銀區與其他需要遠端管理多門市的場景。管理者不必在多個分散視角之間反覆切換，而是能用更直覺的方式理解事件發生的完整背景。

支援問責機制與更快速的事件回顧

重點摘要

從單店監控延伸到連鎖門市管理

對連鎖零售業而言，收銀作業標準化與門市管理一致性非常重要。不同門市可能有不同人流、空間配置、班別安排與交易量，但總部仍需要用一致的方式掌握關鍵作業流程。

360° 全景視覺可以協助總部或區域管理者遠端查看門市現場，理解各店營運狀況。當同一套視覺與 AI 偵測架構能延伸到多個門市時，管理者就能更有效率地比較、回顧與優化作業流程。

未來，類似架構也可延伸到更多智慧零售應用，例如顧客動線分析、異常停留偵測、貨架區域觀察、門市人流統計，或其他與 AI partner 整合的零售場景。

台鍍科技 × Cupola360：AI 工廠安全與全景視覺應用案例

Fri, 28 Aug 2026 03:45:04 GMT

透過全景視覺、PPE 偵測與製程資訊整合，協助高風險製造場域提升安全合規、現場掌握與營運判斷效率。

台鍍科技所處的製造環境，對人員安全、製程可視性與生產控制都有高度要求。熱浸鍍鋅與金屬表面處理製程涉及重型物料、吊掛搬運、製程槽體、溫度控制與嚴格的工安管理，現場安全與營運掌握都不能只仰賴單點影像或事後紀錄。

為了提升場域可視性並支援更安全的現場作業，台鍍科技於工廠關鍵區域部署 16 支 Cupola360 RX1000P 全景相機，並結合 Synology NAS 與第三方 AI 偵測，支援大範圍監控、安全辨識與製程狀態掌握。原文也明確列出，本案部署 16 支 RX1000P，並整合 Synology NAS 與第三方 AI 偵測。

這個案例展現了全景視覺在高風險工業環境中的實際價值：它不只是提供監控畫面，而是協助團隊在同一個視覺脈絡中理解人員、設備、物料與製程條件之間的關係。

挑戰：高風險製造現場需要更完整的安全與製程視野

在熱浸鍍鋅與金屬加工環境中，安全風險可能同時來自多個方向。作業人員會在產線與製程區域中移動，物料需要吊掛與搬運，製程槽體也必須維持特定操作條件。

傳統固定視角相機可以記錄特定角度，例如出入口、單一產線或局部作業區。但當管理團隊需要確認安全規範是否被遵守、回顧事件發生過程，或理解吊掛作業與人員動線的關係時，分散的固定畫面往往不夠完整。

對這類工業場域而言，監控不能只停留在被動錄影。工廠需要的是能夠提升現場可視性、支援工安合規，並把影像資訊與製程條件連接起來的視覺基礎。

解決方案：結合 NAS 整合的 RX1000P 全景攝影機

Cupola360 RX1000P 全景相機部署於台鍍科技工廠內多個關鍵區域，用於監控出入口、生產區、製程相關空間與人員活動區域。透過 360° 全景覆蓋，每支相機都能提供更寬廣的現場視野，相較於傳統固定角度畫面，更有助於降低監控盲區。

影像也透過 Synology NAS 進行儲存與管理，為工廠監控與事件回顧建立穩定基礎。這樣的架構讓台鍍科技能保留完整的全景影像紀錄，同時也支援後續與 AI 偵測、製程資訊或其他營運系統整合。原文說明，RX1000P 相機與 Synology NAS 連接，用於影像儲存與管理，並作為工廠監控與事件回顧的基礎。

結合第三方 AI 偵測，支援安全辨識

吊掛、物料與製程區域監控

從工廠監控走向數位轉型

說明

重點摘要

高雄交通監控與 360° 全景相機：地圖式城市場域感知應用

Fri, 28 Aug 2026 03:39:48 GMT

360° 全景相機可作為地圖式交通平台中的視覺層，與傳統道路攝影機、交通資訊與事件資料共同提供更完整的城市現場脈絡。

高雄是台灣南部的主要大城市之一，擁有繁忙路口、公共運輸路線、觀光景點、停車場域與高密度都市交通活動。對城市管理而言，即時掌握道路狀況不只是看見車流，也包含理解事件、地點與周邊環境之間的關係。

為支援即時交通資訊服務，高雄市提供地圖式交通資訊平台，整合道路順暢度、交通事件、停車資訊、車輛偵測器資料、傳統路口影像，以及 360° 全景相機視角。透過地圖介面，使用者可以從自己關注的位置出發，查看交通狀態，並在需要更多現場脈絡時開啟對應影像。

在這個應用參考中，Cupola360 分享的是 360° 全景相機如何作為地圖式交通平台中的一個視覺層，與既有交通資訊與傳統攝影機畫面搭配使用。它的目的不是取代傳統攝影機，也不是取代既有城市平台，而是透過更寬廣的現場視角，補足固定視角影像較難呈現的周邊脈絡。

挑戰：交通監控需要細節，也需要現場脈絡

城市交通環境相當複雜。一個路口可能同時有來自多方向的車流、機車、巴士、行人、停車活動、鄰近公共設施與突發事件。若只看單一方向的路口影像，往往無法理解完整狀況。

傳統道路攝影機能提供特定方向的清楚細節，例如某一條車道、某個轉彎方向或特定道路段。但壅塞或事件的原因，可能來自旁邊道路、鄰近活動、人流變化、轉向車道、違停、施工，或其他不在該鏡頭視角內的因素。

因此，對城市監控而言，系統不只需要「看得清楚」，也需要「看得完整」。傳統攝影機提供細節，360° 全景相機則能提供更完整的周邊脈絡，協助使用者從單一角度判讀，進一步理解路口或公共空間的整體狀況。

應用方案：地圖式交通平台中的 360° 視角整合

註：本畫面截圖翻譯改編自高雄市即時交通資訊平台，翻譯內容僅供示意參考。原始平台網址： https://traffic.tbkc.gov.tw/

為什麼 360° 全景相機對交通監控有價值

支援既有基礎建設，而不是取代它

說明

重點摘要

Cupola360 x 鴻佰科技 : AI Server Lighthouse Factory 全景視覺管理案例

Fri, 28 Aug 2026 03:21:53 GMT

將工廠可視性、AI 分析與生產資料整合在一起，打造更直覺、更智慧的營運管理方式。

鴻佰科技（Ingrasys）隸屬於鴻海科技集團，專注於 AI computing solutions 與 cloud infrastructure。在南青 AI Server Lighthouse Factory 中，Ingrasys 希望以更智慧的方式監控工廠營運、串接生產資料、支援 AI 分析，並透過更沉浸式的視覺體驗呈現工廠現場。

對智慧製造而言，影像不只是紀錄工具，而是理解現場、連接資料與管理營運的重要入口。Cupola360 透過 720° 全景視覺，協助 Ingrasys 將工廠空間、AI 分析、機台資料、SFC 資訊與即時儀表板整合在同一個更直覺的視覺管理架構中。

挑戰：分散的工廠影像難以形成完整現場理解

在大型製造環境中，傳統 CCTV 系統通常需要大量固定視角相機，才能覆蓋產線、設備區、物料動線與關鍵作業區域。然而，即使現場已安裝多支相機，管理者仍常常需要在不同影像畫面之間來回切換，才能拼湊出現場正在發生什麼事。

這種分散式視角會讓工廠狀態與即時生產資料之間的連結變得不夠直覺。對智慧製造團隊來說，單純錄影已經不夠。他們需要的是一種能同時看見現場、理解資料，並支援遠端營運管理的方式。

解決方案：以 RX1000F 建立 720° 全景覆蓋

為提升可視化程度，鴻佰科技在工廠內部署超過 50支 Cupola360 RX1000F 720° 全景相機。透過統一的全景空間視角，每支相機都能覆蓋更大的區域，並提供更完整的現場脈絡。

根據南青工廠部署經驗，Cupola360 協助相機部署數量減少超過 60%，同時將 AI 處理需求降低超過 50%。在更少相機數量與更完整視覺輸入的基礎上，系統不只提升覆蓋效率，也讓 AI 分析更有效率。

這對 AI Server 工廠特別重要。生產現場通常同時存在高度密集的設備、人員、物料與流程資訊。全景視覺能讓管理者用更少的切換、更大的畫面脈絡，掌握不同區域之間的關係。

串接工廠資料與即時儀表板

支援沉浸式工廠導覽與遠端管理

AI 偵測、區域警示與事件回顧

支援智慧製造的開放式平台

重點摘要

校園安全監控：AI 偵測與全景視覺如何支援更安全的學校環境

Fri, 28 Aug 2026 03:15:57 GMT

透過 AI 偵測與 360° 全景視覺，協助學校更早掌握潛在風險、降低監控盲區，並以更完整的現場脈絡進行應變。

學校本應是學習、成長與交流的空間。然而，現代校園同時也是高度開放且複雜的公共場域。學生、教職員、訪客、家長、合作廠商與外部人員可能在同一天進出校園，不同區域也可能承載教學、活動、行政、運動與公共服務等多種功能。

在這樣的環境中，校園安全不只是出入口管制的問題，也不只是增加更多監視器就能解決。學校需要的是更完整的場域感知：能夠在風險剛出現時更早掌握，在事件發生時更快定位，在事後也能以更清楚的影像脈絡進行回顧。

AI 偵測與 360° 全景視覺，正逐漸成為校園安全監控中的重要輔助工具。它們不應被視為取代人員管理或校園安全制度的方式，而是協助學校更有效理解公共區域狀況、降低盲區，並支援更快速的現場應變。

校園的安全挑戰通常不是單點問題，而是跨區域、跨時間、跨人流的動態問題。

一個事件可能從校門口開始，延伸到走廊、操場、停車場或教學大樓周邊。人員可能在不同區域之間移動，衝突也可能在短時間內升高。若監控系統只能提供零散的固定視角，管理人員就需要不斷切換畫面，才能拼湊出完整現場狀況。

傳統固定式相機對特定方向很有用，例如出入口、電梯口、走廊端點或停車場入口。但在開放場域中，單一固定視角容易留下盲區，也可能讓事件發生位置、移動方向與周邊脈絡變得難以判讀。

360° 全景視覺的價值在於，能從單一視點提供更完整的周邊視野，協助安全人員以更大的場景理解事件，而不是只看到某個局部角度。

為什麼校園安全需要完整場域視野

AI 偵測如何支援更快的校園應變

AI 偵測可以協助校園安全團隊在大量影像中更快發現異常狀況。依照不同場域需求，AI 可用於偵測特定行為、物件或區域事件，例如人員聚集、入侵管制區、異常停留、危險物品出現，或其他需要即時關注的情況。

在校園場景中，AI 的價值不是「自動判斷一切」，而是協助人員更快注意到可能需要處理的事件。

例如，若系統偵測到疑似危險物品，或有人進入限制區域，管理人員可以更快收到警示，並查看對應畫面確認狀況。若警示能與全景視覺結合，操作人員不只知道「系統偵測到什麼」，也能理解事件發生在哪裡、附近是否有人、周圍環境是否需要立即介入。

這種結合能讓 AI 從單純的偵測結果，轉化為更具現場脈絡的安全資訊。

支援事件回顧與責任釐清

校園安全系統的價值不只在於即時應變，也包含事後回顧。

當校園中發生衝突、意外、財物損失、出入爭議或其他需要釐清的事件時，管理單位通常需要回看影像，理解事件發生過程、相關人員動線，以及現場是否有其他背景因素。

如果影像只來自多個分散的固定角度，事件回顧可能需要花費大量時間比對不同畫面。若缺少足夠的場景脈絡，也可能造成判讀困難。

360° 全景影像可提供更完整的現場記錄，協助管理人員以較連續的視角理解事件周邊環境。搭配 AI 偵測或事件標記，也能讓後續查找與回顧更有效率。

這對學校而言很重要，因為校園安全不只是防止事件發生，也包含在事件發生後，以更透明、更有依據的方式進行後續處理。

在安全與隱私之間取得平衡

談到校園安全監控，隱私一定是不能忽略的議題。

學校不是一般商業空間，學生與教職員的日常活動需要受到合理保護。任何 AI 偵測或影像監控系統，都應該以明確目的、必要範圍與適當權限管理為前提。

這代表系統設計時，應優先聚焦於公共區域與安全風險管理，而不是用來追蹤個人日常行為。學校也應明確定義哪些區域需要監控、哪些事件需要警示、誰能存取影像、資料保留多久，以及如何避免不必要的過度監控。

AI 與全景視覺的角色，應該是協助安全人員更快掌握公共區域的風險，而不是取代校園治理、輔導機制或人員判斷。

在導入這類系統時，技術能力與治理原則必須同時被考量。安全與隱私不是二選一，而是需要透過合理設計取得平衡。

360° 全景視覺如何支援更安全的校園環境

學校必須是學生、教師與職員的安全場所。隨著校園環境日趨複雜,安全監控應超越片段式的攝影機畫面與被動式錄影。

AI 偵測與 360° 全景視覺能協助學校及早辨識風險、更快理解事件,並以更完整的情境做出應變。同時,這類系統的部署也必須具備明確的隱私界線、負責任的存取控管與透明的政策。

對學校而言,全景攝影機與 AI 可以相輔相成,打造更完整的安全監控方案。全景攝影機提供更寬廣的視覺覆蓋範圍,AI 則能協助標示出需要留意的特定風險。

這樣的結合能在多方面支援校園團隊:

減少公共區域的視覺死角
協助教職員以更完整的情境理解事件
支援更快速的風險警示
改善緊急應變的協調效率
為事後審查提供更清晰的紀錄
協助學校規劃更完善的監控覆蓋範圍

這對大型校園、開放式校區、公共設施或多建築物環境尤其有幫助。當安全團隊需要遠端查看現場、確認警示、協調人員前往處理，或在事後回顧事件時，完整視野都能降低判讀成本。

360° 全景視覺的價值，不只是「看得更廣」，而是讓校園安全監控更接近現場真實脈絡。

結語

AI 偵測與 360° 全景視覺可以協助學校建立更完整的校園安全感知。它們能幫助管理人員更早發現風險、減少監控盲區、加快事件確認，並在事後提供更清楚的影像脈絡。

但技術本身不是完整答案。校園安全仍需要受訓人員、清楚流程、適當權限管理與合理的隱私治理。

對學校而言，真正有價值的監控系統，不是看更多畫面，而是更快理解現場、做出正確反應，並在保護校園成員的前提下，打造更安全、更值得信任的學習環境。

泳池安全監控與 AI 溺水偵測 : 為什麼完整視野很重要

Fri, 28 Aug 2026 02:57:10 GMT

AI 能協助更快發現異常，但前提是系統必須先看得見完整現場。

游泳池是運動、休閒與家庭活動的場所,尤其在夏季更是如此。公共泳池、飯店、學校、運動中心與社區設施都仰賴救生員與安全團隊來確保泳客安全。

然而,游泳池同時也可能是高風險環境。當有人突然感到不適、抽筋、疲勞,或發生溺水緊急狀況時,最初的幾秒鐘至關重要。在人潮擁擠的泳池中,遇險的泳客未必容易被察覺——他們可能不會大聲呼救,而其動作看起來也可能與漂浮、休息、玩耍或憋氣練習相似。

正因如此,許多場館開始採用 AI 溺水偵測與 AI 輔助泳池監控系統。這些技術有助於偵測異常行為,並在異狀發生時通知工作人員。然而,AI 偵測仰賴一項關鍵條件:系統必須能夠清楚看見其應監控的區域。

AI 偵測不是從演算法開始，而是從「看得見」開始。

如果相機無法拍到泳池角落、扶梯周邊、池邊區域，或人群重疊的擁擠區域，AI 就沒有足夠的視覺資料可以分析。這時問題不只是 AI 模型準不準，而是系統一開始是否真的看見了完整現場。

泳池是一個視覺條件非常複雜的環境。水面反光、波紋、潑水、泳道線、光線變化，以及人體局部沒入水中的狀態，都會讓 AI 系統與現場人員更難即時判讀狀況。即使有受訓救生員在場，也很難在每一秒都同時掌握泳池每個角落。

因此，在評估 AI 泳池安全監控時，不能只看 AI 模型的偵測能力，也必須回到更基礎的問題：相機是否提供足夠的視野？場域覆蓋是否完整？高風險區域是否被清楚看見？

為什麼可視性在 AI 泳池監控中如此重要

泳池安全監控中的常見盲區

視覺死角是 AI 輔助泳池安全中最大的挑戰之一。死角不一定只是攝影機畫面之外的物理區域,在真實泳池環境中,死角可能以多種形式出現。

幾何死角 : 發生於角落、池邊、梯子、兒童區或深水區落在攝影機視野範圍之外時。
遮蔽死角 : 發生於泳客被其他人、泳道分隔線、柱子、池體結構或設備擋住時。
視覺複雜度： 也會影響偵測效果。水面反光、水花、漣漪與部分沒入水中,都可能讓泳客難以被清楚辨識。
行為模糊性： 是另一項挑戰。在泳池中,異常行為可能與正常游泳、漂浮、休息或玩耍的動作十分相似。
作業流程盲區： 即使在警示發出後仍可能發生。工作人員仍需知道警示來自何處、應由誰負責處理,以及需要多快採取行動。

也因此，泳池安全監控不能只用 AI 偵測準確率來評估。相機位置、視野範圍、覆蓋設計與現場應變流程，都同樣重要。原文也指出，camera placement、field of view、coverage design 與 response workflow 都是關鍵。

導入 AI 溺水偵測前，場館應先確認什麼？

在導入 AI 溺水偵測系統前，場館團隊應先評估所有高風險區域是否都能被清楚看見。這些區域包括泳池角落、池邊、扶梯、入口、兒童區、深水區，以及泳客容易重疊或聚集的位置。

除了「看不看得到」，也要進一步評估系統是否能提供足夠的現場脈絡，讓工作人員快速回應。可靠的 AI 輔助泳池監控系統，應該是協助救生員取得更完整的 situational awareness，而不是讓場館產生錯誤的安全感。

AI 可以協助更快發現潛在風險，但它必須與受訓人員、安全作業流程與明確的應變機制一起運作。AI 不應被視為取代救生員的工具，而應該是輔助現場安全管理的視覺與警示系統。

360° 全景視覺如何支援泳池安全

這正是 360° 全景視覺能發揮價值的地方。

更寬廣的視野有助於減少破碎化監控，並提供更完整的泳池周邊環境理解。相較於只仰賴多支固定方向相機，360° 全景視覺能協助管理團隊以更完整的畫面脈絡檢視現場。

對 AI 輔助泳池監控而言，較完整的可視性可支援更好的覆蓋規劃、事件回溯與人機協作。它不會取代救生員、安全程序或人的判斷，但可以協助團隊找出覆蓋缺口、理解事件發生位置，並改善監控系統的部署方式。

換句話說，360° 全景視覺的價值不只是「看得更廣」，而是讓 AI 偵測與現場人員都能在更完整的場域脈絡下工作。

結語

AI 溺水偵測可以成為泳池安全監控的重要輔助工具，但不應被視為單獨運作的完整解決方案。它的效果取決於相機系統能看見什麼，也取決於整個監控與應變流程如何設計。

對公共泳池、飯店、學校、運動中心與休閒設施而言，安全監控應建立在三個基礎上：完整可視性、現場脈絡，以及人機協作。

在 AI 能夠偵測危險之前，系統必須先看見完整現場。

永進機械×Cupola360：全景智慧巡檢，打造更安全的製造現場

Fri, 28 Aug 2026 02:54:07 GMT

透過全景智慧巡檢，串聯工廠視覺監控、安全偵測與工作回報系統。

YCM（永進機械）是一家總部位於台灣的工具機製造商。為支援智慧製造與職場安全管理， 永進機械 攜手 Cupola360 與 DotZero，導入全景智慧巡檢系統。

在此次部署中，Cupola360 提供全景視覺層，DotZero 則負責智慧製造系統整合與 AI 賦能的工作流程應用。兩者結合，協助永進機械提升即時生產可視化程度、安全意識，以及數位化工廠營運能力。

挑戰：串聯生產可視化與安全管理

在製造業環境中，生產管理與職場安全往往是由各自獨立的系統來處理。工廠團隊可能需要監控生產狀態、確認工作人員是否遵守安全規範、應對異常狀況，並透過不同工具或人工流程記錄工作進度。

當現場範圍龐大、工作流程複雜時，固定式攝影機視角與人工回報方式，往往無法提供足夠的可視性或反應速度。團隊需要一種更具連結性的方式，來掌握現場狀況、理解正在發生的事，並在安全或生產問題出現時迅速應對。

解決方案：結合合作夥伴整合的全景智慧巡檢

Cupola360 RX1000F 全景攝影機，有助於提供橫跨整個工廠環境的更廣闊現場可視範圍。團隊不必再仰賴狹窄的攝影機視角，而能以更完整的視覺情境，檢視生產區域，並更深入理解人員、設備與工作流程在現場的互動情形。

透過 DotZero 的智慧製造整合能力，全景視覺層得以與 AI 賦能的應用程式及工廠工作流程相互連結。這使得系統能以更整合的方式，支援即時生產追蹤、安全辨識、事件警示與工作回報。

即時生產可視化

個人防護裝備（PPE）辨識與安全警示

手勢辨識工作回報

支援智慧製造轉型

重點摘要

日本熊害與 360° AI 視覺｜熊害偵測與戶外安全應用

Fri, 28 Aug 2026 02:51:25 GMT

透過 360° 全景相機與 AI 偵測，協助戶外場域更早發現風險、降低監控盲區，並以更完整的現場脈絡進行應變。

近年來，日本熊出沒與人熊遭遇事件逐漸成為公共安全議題。風險不再只發生在山區，也可能出現在住宅周邊、工廠外圍、觀光景點與偏遠設施附近。當野生動物活動範圍越來越接近人類生活與營運場域，及早偵測與更完整的現場掌握就變得更加重要。

對戶外場域管理者而言，挑戰不只是「熊出現後要不要偵測得到」。更關鍵的是，系統是否能幫助人員理解完整現場：風險從哪個方向接近、與人員或建物的距離有多近、現場狀況是否正在變化，以及是否需要即時通知相關人員採取行動。

在大阪近期舉辦的一場展覽中，Cupola360 的日本合作夥伴 iiTEC Co., Ltd. 展示了一項結合 Cupola360 360° 攝影機與 AI 技術的應用方案，用以支援即時熊出沒偵測與現場監控。這個案例展現了全景視覺技術如何應用於更廣泛的野生動物安全與戶外監控領域。

野生動物安全，已成為許多戶外與半偏遠環境所面臨的重要議題。鄰近山區的住宅區、鄰近自然環境的工廠、觀光景點、公園、露營地與公共設施，都可能面臨突如其來的野生動物活動。

在這些環境中，安全團隊往往需要監控廣闊且開放的區域。風險可能從任何方向出現、快速移動，或是從傳統攝影機角度無法涵蓋的位置進入視野。

對於熊的偵測而言，這一點尤其重要。及早察覺能為營運團隊爭取更多時間，通知現場人員、引導民眾遠離該區域，並協調出更安全的應對方式。

挑戰：野生動物風險逐漸靠近人類活動場域

為什麼固定視角相機容易留下戶外監控盲區

傳統相機適合針對特定方向進行重點監控，例如出入口、道路、圍籬或單一作業區域。但在戶外環境中，風險往往不會只從固定方向出現。

樹木、建物、車輛、地形高低差或相機視角之間的縫隙，都可能形成監控盲區。若要覆蓋較大的場域，管理者通常需要安裝多支相機，並同時查看多個影像來源。即使如此，操作人員仍可能需要頻繁切換畫面，才能理解完整現場狀況。

對戶外安全而言，分散的視角會拖慢判斷速度。當潛在風險出現時，操作人員需要快速掌握完整場景，而不只是看到某一個方向的局部畫面。

應用方式：360° 相機結合 AI 熊害偵測

360° 全景攝影機能夠從單一定點，提供完整的周邊環境視野。當結合 AI 偵測技術後，系統便能持續分析更廣泛的場景，協助更早期地識別潛在風險。

在熊隻偵測情境中，AI 可設定為偵測特定區域內的特定物體或動態。當系統識別出潛在風險時，便能即時向現場人員發送警示。

這類應用能夠協助支援：

在指定戶外區域進行熊隻偵測
向現場人員發送即時警示
完整場景影像紀錄，供事後檢視事件
更廣泛的現場監控意識，減少監控死角
針對設施、觀光景點與公共空間的遠端監控

註：此為示意用途之概念圖，僅供展示參考，並非實際部署畫面或經驗證之偵測結果。

360° 視覺技術的價值，不僅在於更廣的涵蓋範圍，更在於協助營運人員理解事件發生的完整情境脈絡。

在戶外環境中，情境脈絡至關重要。營運人員可能需要知道：動物從何處進入、附近是否有人員活動、風險與建築物或工作區域的距離有多近，以及情況是正朝著關鍵區域移動，還是逐漸遠離。

全景視角能在單一連續畫面中，提供這樣的完整情境資訊。營運人員無需在多個彼此獨立的攝影機角度之間切換，便能更直覺地掌握周邊環境，並做出更快速的判斷。

這對於人力無法持續、全方位監控每個方向的場域尤其實用，例如偏遠設施、山區、鄰近自然環境的工廠、公園、觀光景點與公共空間。

不僅止於熊隻偵測：更廣泛的戶外安全應用

儘管此案例聚焦於日本的熊隻偵測應用，但相同的 360° AI 視覺架構，同樣能支援許多其他戶外安全應用情境。

搭配合適的 AI 模型與系統整合，全景視覺資料可應用於：

野生動物監控
入侵偵測
周界安全防護
工廠邊界監控
倉儲與物流場地監控
智慧城市公共安全
觀光景點監控
偏遠設施安全意識

上述多數應用，皆共享同一項核心挑戰：團隊需要更早期的偵測、更少的監控死角，以及更完善的現場情境掌握能力，才能有效展開應變行動。

Cupola360 的角色：提供 AI 合作夥伴的 360° 視覺平台

在由合作夥伴主導的 AI 應用方案中，Cupola360 提供 360° 全景攝影機與視覺平台，作為 AI 合作夥伴、系統整合商與解決方案供應商得以在其上進一步開發的基礎。

一台 360° 攝影機的價值，遠不止於「攝影機」本身。它能作為 AI 應用的視覺基礎，提供廣域全景資料，支援偵測、警示、監控與事件回顧等功能。

對於野生動物偵測與其他戶外安全應用而言，這意味著合作夥伴能將自身的 AI 技術能力，與 Cupola360 的全景視覺相結合，針對特定產業、場域與營運環境，打造專屬的解決方案。

Cupola360 樂見如 iiTEC 這樣的合作夥伴，將創新的 360° AI 視覺應用帶入實際場域。隨著戶外安全挑戰持續演變，全景視覺技術將有助於打造更具彈性、更能掌握情境脈絡的跨產業監控系統。

為什麼 360° 全景視覺有助於戶外場域感知

如何快速計算 360° 全景相機的有效像素？

Mon, 10 Aug 2026 05:45:45 GMT

在規劃智慧工廠、智慧建築、資料中心或機器人巡檢等專案時，客戶經常會問一個很實際的問題：360° 全景相機到底可以看多遠？更具體來說，問題通常會是：

360° 全景相機最遠可以偵測到人嗎？
相機應該安裝在哪個位置最合適？
目前的解析度是否足夠支援 AI 偵測？
如果要辨識特定物件，距離應該如何規劃？

這些問題不能只用「相機解析度」或「可視距離」來回答。真正影響 AI 偵測與視覺判讀效果的，是目標物在影像中實際佔了多少像素。也就是有效像素。

在 360° 全景影像應用中，有效像素可以作為專案初期部署評估的快速估算方法。只要知道相機解析度、目標物實際尺寸，以及相機與目標物之間的距離，就能初步判斷該相機是否適合特定應用場景。

為什麼需要計算有效像素？

在監控系統規劃中，許多人會先比較相機規格，例如 2MP、4K、8MP 或更高解析度。然而，解析度本身並不能直接代表實際偵測效果。

同樣是 4K 影像，如果目標物距離相機很遠，它在畫面中可能只佔十幾個像素；如果目標物距離較近，即使相機解析度相同，它在畫面中的像素數也會明顯增加。

對 AI 偵測而言，模型並不是直接判斷「距離」，而是分析影像中的特徵。當目標物在畫面中太小、像素不足、邊緣不清楚或細節不足時，偵測效果就會下降。這也是為什麼在評估相機安裝位置時，不能只問「可以看多遠」，而要進一步估算目標物在畫面中有多少有效像素。

有效像素可以協助系統整合商與客戶更客觀地評估：

相機安裝距離是否合適
目標物在畫面中是否足夠清楚
AI 偵測是否具備基本影像條件
是否需要提高解析度、縮短距離或增加相機數量
是否應搭配專用相機，例如 LPR 車牌辨識相機

360° 全景影像的估算邏輯

對 equirectangular 360° 全景影像而言，水平解析度代表完整的 360° 視角。也就是說，影像的水平像素會分布在相機周圍一整圈。

因此，可以將相機到目標物的距離視為圓的半徑，而目標物的實際寬度則可以視為它在圓周上所佔的一小段長度。目標物在圓周上佔的比例，約略對應到它在 360° 全景影像水平解析度中所佔的比例。

快速估算公式如下：

Effective Pixel Width = 水平解析度 × 目標物寬度 ÷ (2 × π × 距離)

若以 Cupola360 3840 × 1920 全景相機為例，水平解析度為 3840 pixels，因此公式可寫成：

Effective Pixel Width = 3840 × 目標物寬度 ÷ (2 × π × 距離)

只要知道以下三個條件，即可快速估算目標物在影像中的有效像素寬度：

相機水平解析度
目標物實際寬度
相機與目標物之間的距離

這是一個快速估算方法，適合用於專案初期評估。實際結果仍會受到安裝高度、鏡頭角度、光線、物體姿態、遮擋、影像壓縮、AI 模型與系統設定等因素影響。

範例一：人員偵測

範例二：車牌辨識

以成人肩寬或人體正面寬度約 0.5 公尺作為估算基準，使用 3840 × 1920 的 360° 全景影像時，可得到以下結果：

再以台灣一般小客車車牌寬度約 0.38 公尺作為估算基準，同樣使用 3840 × 1920 的 360° 全景影像，可得到以下結果：

有效像素與 360° 全景相機部署規劃

有效像素的計算可以協助專案團隊在初期快速判斷相機配置是否合理。例如，在智慧工廠中，若目標是偵測人員是否進入危險區域，系統整合商可以先估算人員在不同距離下的有效像素，判斷相機是否需要靠近出入口、產線或安全區域。

在資料中心或設備機房中，若目標是進行走道、人員或設備區域監控，則可透過有效像素估算相機安裝位置，避免目標物在遠端畫面中過小。

在機器人巡檢應用中，360° 全景相機可協助取得周圍環境資訊，而有效像素估算則能協助判斷不同距離下的物件是否具備基本辨識條件。

這個方法特別適合用於：

相機安裝點位初步規劃
AI 偵測距離評估
相機解析度選型
現場測試前的可行性估算
與客戶討論部署範圍與預期效果
判斷是否需要搭配其他專用相機

雖然有效像素是一個實用的估算工具，但它不能取代現場測試。實際偵測效果仍會受到多項因素影響，包括：

相機安裝高度與俯角
目標物是否正對相機
光線條件與夜間照明
動態模糊與快門速度
壓縮率與串流品質
目標物是否被遮擋
背景複雜度
AI 模型類型與訓練資料
偵測任務是人員、車輛、PPE、車牌或其他物件

因此，有效像素的價值在於快速建立初步判斷。它能幫助團隊在專案早期縮小設計範圍，找出較合理的安裝位置與相機規格，再搭配現場測試進行驗證。

Supporting AI-Enabled Monitoring Across Different Sites

注意事項：有效像素不是唯一判斷標準

結論

有效像素是一個簡單但實用的技術概念。透過目標物尺寸、相機水平解析度與相機距離，即可快速估算目標物在 360° 全景影像中的大致像素寬度。

對智慧工廠、智慧建築、資料中心、公共空間與機器人巡檢等應用而言，這個方法能協助系統整合商與客戶更有效率地進行相機選型與部署規劃。

在專案初期，與其只問「360° 全景相機可以看多遠」，更精準的問題應該是：目標物在這個距離下，是否仍有足夠的有效像素支援偵測與判讀？

未來，我們也將持續分享更多 360° 全景視覺的技術知識，包括 DORI、Pixels Per Meter（PPM）、AI 偵測距離與全景相機安裝設計，協助合作夥伴與客戶更深入理解 360° 視覺技術在不同場域中的應用。

從估算結果可以看出，距離越遠，人體在影像中的有效像素寬度會快速下降。在實務應用中，3840 × 1920 全景相機約可在 10 至 15 公尺範圍內提供較穩定的人員偵測條件。

不過，這並不代表超過 15 公尺就完全無法偵測，也不代表 10 公尺內一定能在所有情境下穩定偵測。實際效果仍會受到安裝高度、人物姿態、移動速度、光線、遮擋、背景複雜度與 AI 模型訓練資料影響。

因此，有效像素應被視為部署評估的第一層判斷，而不是唯一判斷標準。

車牌辨識通常需要更高的像素密度與更穩定的拍攝條件，包括適當角度、足夠解析度、快門速度、補光、低壓縮影像，以及針對車牌辨識設計的演算法。

因此，360° 全景相機較適合作為全域監控、事件掌握與車輛動線追蹤的視覺基礎。若專案目標是高成功率的車牌辨識，仍建議搭配專用 LPR 相機，並將 360° 全景相機作為場域全貌與事件脈絡的補充。

透過目標物尺寸、相機解析度與安裝距離，快速估算 360° 全景相機在不同場域中的可視範圍與 AI 偵測適用性。

How to Quickly Estimate Effective Pixels for 360° Panoramic Cameras

Mon, 10 Aug 2026 05:39:16 GMT

When planning smart factory, smart building, data center, or robotic inspection projects, one of the most common questions is: how far can a 360° panoramic camera see?

In practical discussions, this question often becomes more specific:

Can the camera detect a person at a certain distance?
Where should the camera be installed?
Is the current resolution enough for AI detection?
How far can the system identify or track a target object?

These questions cannot be answered by resolution alone. What matters most for AI detection and visual verification is how many pixels the target object occupies in the image.

This is where effective pixels become important.

For 360° panoramic vision, effective pixel estimation provides a simple way to evaluate whether a camera setup is suitable for a specific application. By knowing the camera resolution, the real-world size of the target object, and the distance between the camera and the object, project teams can quickly estimate whether the object will appear large enough in the image for detection or review.

A practical method for estimating whether a 360° panoramic camera has enough pixel coverage for AI detection, remote monitoring, and deployment planning.

In video surveillance and monitoring projects, camera specifications are often compared by resolution: 2MP, 4K, 8MP, or higher. However, resolution alone does not determine detection performance.

With the same 4K image, a target object far away from the camera may occupy only a small number of pixels. If the same object is closer to the camera, it will occupy more pixels and provide more visual detail.

AI models do not directly analyze “distance.” They analyze image features. If the target object is too small in the image, lacks edge definition, or does not contain enough detail, detection performance may decrease.

For this reason, camera planning should not only ask “how far can the camera see?” A more useful question is: how many effective pixels does the target object have at that distance?

Effective pixel estimation can help system integrators and customers evaluate:

Whether the camera installation distance is suitable
Whether the object appears large enough in the image
Whether the image provides a basic condition for AI detection
Whether higher resolution, shorter distance, or additional cameras are needed
Whether a dedicated camera, such as an LPR camera, should be used for specific tasks

Estimation Logic for 360° Panoramic Images

For an equirectangular 360° panoramic image, the horizontal resolution represents the full 360° field of view. In other words, the horizontal pixels are distributed around the entire circle surrounding the camera.

A simple way to estimate effective pixel width is to treat the distance from the camera to the object as the radius of a circle. The real-world width of the object occupies a portion of the circle’s circumference. That same proportion can be applied to the horizontal resolution of the panoramic image.

The quick estimation formula is:

Effective Pixel Width = Horizontal Resolution × Object Width ÷ (2 × π × Distance)

For a Cupola360 3840 × 1920 panoramic camera, the horizontal resolution is 3840 pixels. The formula becomes:

Effective Pixel Width = 3840 × Object Width ÷ (2 × π × Distance)

To use this estimation, you need three values:

Camera horizontal resolution
Real-world object width
Distance between the camera and the object

This is a quick estimation method for early-stage project planning. Actual performance may still be affected by installation height, camera angle, lighting, object pose, occlusion, compression, AI model capability, and system configuration.

Example 1: Person Detection

Example 2: License Plate Recognition

Assume the approximate width of an adult person is 0.5 meters. With a 3840 × 1920 360° panoramic image, the estimated effective pixel width is:

For a standard passenger vehicle license plate in Taiwan, assume the plate width is approximately 0.38 meters. With the same 3840 × 1920 360° panoramic image, the estimated effective pixel width is:

Using Effective Pixels for Camera Deployment Planning

Effective pixel estimation can help project teams quickly evaluate whether a proposed camera layout is reasonable.

In a smart factory, for example, if the goal is to detect whether a person enters a restricted or hazardous area, system integrators can estimate the effective pixel width of a person at different distances. This helps determine whether the camera should be installed closer to entrances, production lines, or safety zones.

In data centers and equipment rooms, if the goal is to monitor aisles, personnel movement, or equipment zones, effective pixel estimation can help avoid placing the camera too far from critical areas.

In robotic inspection applications, 360° panoramic cameras can provide surrounding visual awareness, while effective pixel estimation helps determine whether objects at different distances have enough image detail for detection or review.

This method is useful for:

Initial camera placement planning
AI detection distance estimation
Camera resolution selection
Feasibility checks before site testing
Discussion of expected coverage with customers
Deciding whether additional dedicated cameras are needed

Effective pixels provide a practical estimation, but they do not replace on-site testing.

Actual detection performance may be affected by many factors, including:

Camera installation height and tilt angle
Whether the object faces the camera
Lighting conditions and night illumination
Motion blur and shutter speed
Compression level and streaming quality
Occlusion
Background complexity
AI model type and training data
Detection target, such as people, vehicles, PPE, license plates, or other objects

The value of effective pixel estimation is that it provides a first-level technical reference. It helps teams narrow down deployment options, identify reasonable camera positions, and select suitable camera specifications before validating the design on site.

Supporting AI-Enabled Monitoring Across Different Sites

Important Considerations

Conclusion

Effective pixel estimation is a simple but useful concept. By using object size, camera horizontal resolution, and distance, project teams can quickly estimate the approximate pixel width of a target object in a 360° panoramic image.

For smart factories, smart buildings, data centers, public spaces, and robotic inspection applications, this method helps system integrators and customers plan camera selection and deployment more efficiently.

Instead of asking only “how far can a 360° panoramic camera see?”, a more accurate question is: at this distance, does the target object still have enough effective pixels for detection and visual review?

In future technical articles, we will continue to share more topics related to 360° panoramic vision, including DORI, Pixels Per Meter (PPM), AI detection distance, and panoramic camera installation design.

The result shows that effective pixel width decreases quickly as distance increases. In practical applications, a 3840 × 1920 panoramic camera can generally provide more stable conditions for person detection within approximately 10 to 15 meters.

This does not mean detection becomes impossible beyond 15 meters, nor does it guarantee perfect detection within 10 meters. Actual results still depend on installation height, human posture, movement speed, lighting, occlusion, background complexity, and AI model design.

Effective pixel estimation should be used as an initial planning reference, not as the only performance criterion.

License plate recognition usually requires higher pixel density and more controlled imaging conditions, including proper angle, sufficient resolution, shutter speed, illumination, low compression, and an algorithm designed specifically for LPR.

For this reason, 360° panoramic cameras are better suited for full-scene monitoring, event awareness, and vehicle movement context. If the project requires high-success-rate license plate recognition, a dedicated LPR camera is still recommended. The 360° panoramic camera can then provide the surrounding scene context and event-level awareness.

Why Effective Pixels Matter

DP Smart × Cupola360: Opening a 360° Live View of Smart Cities and Public Spaces

Tue, 04 Aug 2026 09:07:57 GMT

How scalable 360° live streaming can connect tourism sites, public spaces, traffic points, events, venues, and community activities into a real-time visual network.

DP Smart is one of Cupola360’s key partners in Taiwan, with extensive experience in deploying 360° live image applications across public-facing and remote monitoring environments.

Through the Rogy360 / 360 See Taiwan live image network, DP Smart demonstrates how 360° live streaming can go beyond a single scenic webcam. The model can support national parks, scenic destinations, old streets, temples, urban landmarks, zoos, traffic points, public events, exhibitions, venues, and community activities. Instead of serving only one site or one camera, 360° live streaming becomes a scalable visual network for public spaces and smart city applications.

This partner application shows how panoramic live video can help cities, venues, tourism operators, and public-space managers provide real-time visual access to locations that people care about. It also shows how the same 360° video source can support public communication, remote awareness, digital engagement, and management-side integration.

The Challenge: Public Spaces Need Real-Time Visual Access

Public spaces are dynamic. Weather changes, crowds gather, traffic conditions shift, events begin and end, and site conditions may change within minutes.

For visitors, citizens, event participants, or remote viewers, static information is often not enough. A photo may show what a place looks like, but it cannot show what is happening there now. Written updates may provide basic information, but they cannot fully communicate atmosphere, visibility, crowd levels, road conditions, or event activity.

For operators and public-space managers, the challenge is also operational. Many public environments are distributed across different locations, such as scenic viewpoints, transportation nodes, road intersections, venues, parks, visitor centers, and event spaces. Monitoring these places manually can be costly, slow, and fragmented.

360° live streaming provides a practical way to make these spaces visible in real time.

From Scenic Streaming to a Public Live View Network

DP Smart’s 360° live streaming model is not limited to tourism. While scenic destinations are an important application, the broader value is the ability to create a public live view network across many types of open environments.

A single 360° live camera can show one location. A network of 360° live cameras can help the public and site operators understand multiple places over time.

This creates a new kind of real-time visual layer for public spaces. Viewers can open a live stream, look around the scene, and understand the current environment without being physically present. Operators can also use the same visual source as a foundation for remote awareness, event review, or future AI-enabled applications.

In this model, live streaming is not only a media tool. It becomes part of a smart city visibility infrastructure.

Field Categories: Where 360° Live Streaming Can Be Applied

Why YouTube Live Works for Public Viewing

From Public Distribution to Management-Side Monitoring

Cupola360’s Role: The Panoramic Visual Foundation

A Broader View of Smart City Visibility

Related Resources

Based on DP Smart’s pu blic live image applications, 360° live streaming can suppo rt a wide range of field categories.

Tourism & Scenic Destinations

National parks, scenic areas, mountain viewpoints, coastlines, visitor centers, and tourism routes can use 360° live streaming to provide real-time scenery, weather visibility, crowd awareness, and digital destination engagement.

For visitors, this creates a more immersive way to preview a location before arrival. For destination operators, it turns live scenery into an always-on communication channel.

Urban Landmarks & Public Spaces

Old streets, temples, city landmarks, plazas, and high-traffic public areas can use 360° live views to connect online audiences with the atmosphere of a place.

This is valuable for city branding, tourism promotion, public communication, and remote access to culturally significant locations.

Traffic, Roads & Public Safety

Road sections, intersections, traffic nodes, coastal roads, mountain routes, and public safety points can benefit from wider visual context.

Unlike fixed-view cameras that only show one direction, 360° cameras can provide a broader view of surrounding conditions. This can support traffic awareness, roadside observation, public safety response, and remote field monitoring.

Events, Exhibitions & Venues

Exhibitions, forums, esports events, public gatherings, and venue-based activities can use 360° live streaming to bring online audiences closer to the event environment.

Instead of watching only one directed camera angle, viewers can explore the venue atmosphere and understand the event space more freely. For organizers, this creates a new way to extend event visibility beyond the physical venue.

Sports and Community Activities

Sports events, school activities, community gatherings, and outdoor competitions can use 360° live streaming to provide broader field coverage and a more immersive viewing experience.

This is especially useful for events where activity happens across a wide area, and a single fixed camera cannot capture the full context.

Zoos, Parks, and Education Spaces

Zoos, parks, ecological areas, and education-oriented venues can use 360° live views for public engagement, educational outreach, and long-term content operation.

For example, a live panoramic view can allow viewers to observe animal habitats, park environments, or public education spaces remotely.

For public-facing applications, YouTube Live provides a familiar and low-friction distribution channel. Viewers can open a link, watch from desktop or mobile devices, share the stream, and access live content without entering a closed monitoring system.

This makes YouTube suitable for:

Public scenic views
City and landmark live streams
Event and venue visibility
Tourism and destination marketing
Community activities
Public communication
Open live image networks

In this structure, YouTube serves as the public viewing layer, while the 360° panoramic camera provides the immersive visual source.

YouTube Live and Cupola360 ⁺ Patrol serve different roles in a 360° live streaming architecture.

For public-facing applications, YouTube Live is suitable for large-scale content distribution. It allows viewers to access live streams through familiar links, watch from different devices, share the stream easily, and join without entering a closed monitoring system. This makes it a practical choice for tourism promotion, public scenic views, event exposure, city landmarks, and open-access live image channels.

However, YouTube is mainly a public viewing and content distribution layer. It is not designed to replace a management-side monitoring platform when a project requires controlled access, multi-camera operations, AI event review, map-based navigation, system integration, or operational dashboards.

For these use cases, Cupola360 ⁺ Patrol can serve as the management-side layer. Patrol is better suited for internal monitoring workflows, where operators need to view multiple 360° sources, manage site-based layouts, review events, integrate AI alerts, or embed panoramic views into operational systems.

In short:

YouTube Live is suitable for public viewing, high-volume audience access, sharing, and outreach.
Cupola360 ⁺ Patrol is suitable for management-side monitoring, controlled access, AI/event integration, and operational workflows.

For partners like DP Smart, this creates a flexible deployment model. A project can use YouTube Live to reach the public, while extending the same type of 360° video source into Patrol or other management platforms when operational control is required.

In this partner application model, DP Smart provides project deployment, live image integration, public viewing channels, and site-specific application design. Cupola360 provides the 360° panoramic visual foundation that enables immersive live views and wider field awareness.

The value of Cupola360-powered 360° live streaming is its flexibility. The same panoramic camera feed can support public viewing, website integration, tourism platforms, event live streams, public-space awareness, management dashboards, or AI-enabled monitoring workflows, depending on project requirements.

This flexibility is important for partners building live image services across many field types, from tourism and city landmarks to transportation, events, public safety, and remote sites.

Smart city applications often depend on real-time data. In many cases, visual information is one of the most intuitive forms of real-time data because it helps people understand what is happening in a place immediately.

360° live streaming adds a wider layer of visual context. It allows cities, agencies, venues, and operators to provide real-time access to public environments without limiting viewers to a single fixed angle.

For public audiences, this means a more immersive way to experience places remotely. For operators, it means a scalable visual foundation for remote awareness and future AI applications.

DP Smart’s 360° live image network demonstrates how partners can turn panoramic live streaming into a repeatable model for smart tourism, public spaces, event visibility, and city-scale awareness.

Key Highlights

Demonstrates how 360° live streaming can scale across different public environments
Extends beyond scenic tourism into public spaces, traffic points, events, venues, sports, parks, and community activities
Uses YouTube Live as a familiar public-facing distribution channel
Makes 360° live views easy for visitors, citizens, and online audiences to access
Supports tourism promotion, public communication, event visibility, and remote awareness
Uses Cupola360 panoramic camera series for 360° live panorama as the visual foundation
Can be extended into Cupola360 ⁺ Patrol for management-side monitoring and system integration
Supports a broader smart city visibility model for public spaces and remote sites

Note : Explore more 360° live views from YouTube channel.

DP Smart × Cupola360：以 360° 即時影像打造太魯閣智慧觀光平台

Tue, 04 Aug 2026 03:49:21 GMT

透過沉浸式 360° 即時串流，協助景區提供即時景色資訊、行前預覽體驗與遠端場域掌握。

DP Smart 達標智源是 Cupola360 在台灣的重要合作夥伴之一，長期投入 360° 即時影像應用建置，應用場域涵蓋觀光景點、公共空間與遠端監控環境。

在太魯閣國家公園的應用中，DP Smart 以 Rogy360 Pro powered by Cupola360 建置 720 即時影像平台，將 360° 全景即時影像、4K 視覺品質與戶外場域部署能力，整合進智慧觀光應用之中，服務對象同時包含一般旅客與場域管理單位。

這個合作應用展現了 360° 即時串流不只是單純的 webcam 直播。透過將多個全景即時影像視點整合成地圖式觀光平台，景區能提供更即時的現場資訊，提升旅客互動體驗，並強化大型戶外場域的遠端掌握能力。

挑戰：大型景區需要即時場域可視性

傳統觀光資訊多半仰賴靜態照片、文字公告、天氣資訊或人工更新的旅遊提醒。這些資訊雖然有幫助，卻無法完整呈現旅客規劃行程當下的現場狀況。

對太魯閣這樣的大型戶外景區來說，這項限制更加明顯。太魯閣國家公園面積廣大，橫跨峽谷、河流、道路、山區與多個熱門景點，每年吸引大量旅客造訪。旅客在出發前，可能希望先確認即時景色、天氣能見度、人潮狀況或道路條件；而管理單位也需要更有效率的工具，遠端掌握偏遠地形、協調現場作業，並因應快速變化的戶外環境。

在 DP Smart 的太魯閣專案中，智慧觀光面臨的挑戰包括：旅客無法在抵達前預覽真實景色、即時天氣與人潮資訊不足、偏遠地形使人工巡查成本提高，以及天然災害發生時可能影響反應速度。

解決方案：地圖式 720 即時影像平台

為了支援更智慧化的觀光服務，DP Smart 透過 360° 全景相機，將太魯閣多個重要景點的即時影像串流整合至 720 即時影像平台。

這個平台並不是將每一支直播視為獨立影像，而是將多個 360° 視點整理成地圖式即時影像介面。使用者可以從地圖上理解太魯閣作為一個完整景區的空間關係，並依照不同景點位置進行瀏覽。

旅客可以在出發前先查看不同地點的現場狀況；管理單位則能透過即時影像，掌握重要景點、入口區域與道路周邊的視覺資訊。

太魯閣 720 即時影像平台涵蓋八個主要視點，包括天祥、布洛灣吊橋、布洛灣台地、長春祠、太魯閣遊客中心、太魯閣入口、崇德休憩區與匯德觀景台等地點。

每一個 360° 視點都能提供比傳統固定視角更完整的周邊視野。對景區而言，這特別適合用於觀察人流、道路、天氣、景色與能見度等會隨時間變化的現場狀況。

旅客體驗：出發前先看見真實景色

管理價值：從即時觀看到遠端場域掌握

為什麼 360° 全景即時影像很重要

平台延伸：從 YouTube Live 到管理端整合

未來延伸：AI、安全與永續應用

Related Resources

在對外觀光應用中，360° 全景即時影像能提供更沉浸式的行前預覽體驗。

旅客可以查看當下景色、天氣與能見度，觀察景點是否擁擠，並透過 360° 視角感受現場環境氛圍。相較於靜態照片或一般單向 webcam，360° 全景即時影像能提供更完整的空間脈絡，也讓旅客更接近「身在現場」的感受。

這對景區經營尤其有價值。即時影像不只協助旅客做出更好的行程判斷，也能成為景區對外溝通與目的地行銷的常態化內容渠道。

對太魯閣來說，live image 不只是監看功能，而是數位旅遊體驗的一部分。

同一套 360° 即時串流架構，也能支援景區管理需求。

不同使用者需要不同觀看方式。一般旅客可能偏好透過 YouTube 或觀光網站快速開啟直播畫面；而管理單位則可能需要更偏向營運管理的介面，用於同時監看多個地點、檢視即時狀況，或協調現場應變。

這正是 360° 即時串流在智慧觀光中的重要價值：同一個全景影像來源，可以依照不同情境支援公共溝通、旅客互動與管理端場域掌握。

傳統固定式相機通常只能呈現單一方向，例如橋梁入口、道路路段、觀景平台或遊客中心周邊。這類畫面有其用途，但對大型戶外場域而言，單一方向往往不足以呈現完整環境脈絡。

360° 全景相機則能從單一視點捕捉周圍環境，讓旅客與管理者不只看到某個目標區域，也能理解該區域周邊的整體狀況。

對觀光景區而言，360° 全景即時影像可支援：

旅客出發前的即時景色預覽
景點與公共區域周邊的廣域可視性
天氣、能見度與人潮狀況掌握
道路、入口與遊客服務設施的遠端觀察
更沉浸式的數位觀光體驗
對外溝通與景區行銷的即時內容
未來 AI 監測與安全應用的視覺基礎

在目前的太魯閣公開網站中，旅客可透過地圖介面選擇景點，並開啟對應的 YouTube Live 即時影像。對一般大眾而言，YouTube 具備熟悉、好分享、容易觀看的特性，因此很適合作為公開直播與觀光推廣的入口。

若專案需要進一步支援管理端應用，同一套 360° 即時影像架構也可以延伸至更偏營運管理的工作流程。Cupola360⁺ Patrol 可依專案需求，支援多畫面監看、全景影像瀏覽、地圖或場域式導覽、事件疊圖，以及嵌入式營運儀表板等應用。

這個差異很重要。YouTube 適合公開觀看與對外溝通；Patrol 則更適合管理端監控、內部作業與系統整合。

對智慧觀光專案而言，兩種模式可以並行。景區可透過 YouTube 提供對外即時影像，也可依管理需求，將同一類 360° 影像來源延伸至 Patrol 或其他管理平台，用於內部監看、AI 警示或現場協調。

雖然此案例聚焦於 360° 全景即時影像與智慧觀光，但 DP Smart 的太魯閣平台概念也展現了全景影像在 AI 與安全應用上的延伸潛力。

例如，透過 AI 智慧監測，可進一步支援人流統計、虛擬圍籬與移動軌跡追蹤。這類功能能協助管理者掌握遊客分布、設定限制區域，並分析景區中的人員移動行為。

對大型戶外景區而言，360° 視覺資料也可延伸至災害監測與旅客安全應用，例如落石偵測、水位與雨量監測、救援協調與交通管制。這些應用可協助管理單位在地形複雜、天候變化快速的環境中，取得更即時的現場視覺資訊。

此外，長期累積的全景影像紀錄，也可能成為永續溝通、生態監測與 ESG 相關揭露的輔助素材，例如植被恢復紀錄、生物多樣性觀察、保育成果展示與永續報告內容。

關鍵亮點

為太魯閣智慧觀光打造 720 即時影像平台
採用 Cupola360 全景相機作為 360° 全景即時影像基礎
支援 4K 全景即時影像與戶外場域部署
串流太魯閣八個主要景點的即時影像
透過地點式瀏覽與 YouTube Live 提供公開觀看入口
支援旅客行前預覽、天氣與景色掌握、數位觀光互動
建立可延伸至管理端監控、AI 應用、災害應變與永續應用的視覺基礎
展示 360° 即時串流如何成為智慧觀光平台的一部分

備註 : 此影片展示 DP Smart 打造的太魯閣 720 即時影像網站。旅客可於地圖上選擇不同景點，並開啟對應的 360° 即時影像。點擊下方連結即可前往體驗。

DP Smart × Cupola360 | 360° Live Streaming for Taroko Smart Tourism

Tue, 04 Aug 2026 03:32:01 GMT

How 360° live panorama helps scenic destinations provide real-time visitor information, immersive site previews, and remote field awareness.

DP Smart Technology is one of Cupola360’s key partners in Taiwan, with extensive experience in building 360° live image applications for tourism sites, public spaces, and remote monitoring environments.

For Taroko National Park, DP Smart developed a 720 live image platform using Rogy360 Pro powered by Cupola360. The platform brings 360° live panorama, 4K visual quality, and outdoor-ready deployment into a smart tourism application designed for both public visitors and site management needs.

This partner application shows how 360° live streaming can go beyond a simple webcam. By organizing multiple panoramic live views into a map-based tourism platform, scenic destinations can provide real-time visual information, improve visitor engagement, and support wider field awareness across large outdoor environments.

The Challenge: Scenic Destinations Need Real-Time Field Visibility

Traditional tourism information often relies on static photos, written notices, weather reports, or manually updated announcements. While useful, these formats cannot fully show what is happening on site at the moment visitors are planning their trip.

For a large outdoor destination like Taroko, this limitation becomes more important. Taroko National Park covers 92,000 hectares, receives more than 4 million visitors per year, and spans complex high-mountain terrain, marble gorges, rivers, roads, and scenic viewpoints.

Visitors may want to preview the actual scenery, check weather and visibility, understand crowd levels, or evaluate road conditions before arrival. Site managers may also need better tools to monitor remote terrain, coordinate field operations, and respond to fast-changing outdoor conditions.

In DP Smart’s Taroko project presentation, the digital tourism challenge is summarized through several pain points: visitors cannot preview real scenery before arrival, real-time weather, crowd, and road information may be limited, remote terrain makes manual patrol costly, and response can be slower when natural disasters occur.

The Solution: A Map-Based 720 Live Image Platform

To support smarter tourism services, DP Smart uses 360° panoramic cameras to stream live views from key scenic locations across Taroko. Instead of treating each live stream as an isolated camera feed, DP Smart organizes multiple viewpoints into a map-based live image platform.

This allows users to understand Taroko as a connected scenic environment. Visitors can explore different locations before arrival, while site managers can use real-time visual information to support awareness across important viewpoints and access areas.

The Taroko 720 live image platform includes live views from eight key locations, including Tianxiang Zhihui Bridge, Buluowan Suspension Bridge, Buluowan Terrace, Changchun Shrine, Taroko Visitor Center, Taroko Gorge Entrance, Chongde Rest Area, and Huide Scenic Overlook.

Each 360° viewpoint helps provide more complete surrounding visibility than a fixed-direction camera. This is especially useful in scenic areas where visitors, roads, terrain, weather, and visibility conditions may change throughout the day.

Visitor Experience: Previewing the Destination Before Arrival

Management Value: From Live Views to Remote Site Awareness

Why 360° Live Panorama Matters

Platform Extension: From YouTube Live to Management-Side Integration

Future Expansion: AI, Safety, and Sustainability Applications

Related Resources

For public-facing tourism applications, 360° live panorama creates a more immersive way to preview a destination.

A visitor can check the current scene, observe weather and visibility, see whether a viewpoint appears crowded, and experience the atmosphere of a scenic area through a panoramic live view. Compared with static photos or standard webcams, 360° live panorama provides stronger spatial context and a more realistic sense of place.

This helps scenic destinations provide more useful information before the visitor arrives. It can also support destination marketing by turning real-time views into an always-on visual communication channel.

For Taroko, this means live imagery is not only a monitoring function. It becomes part of the digital visitor experience.

The same 360° live streaming architecture can also support site management. DP Smart’s Taroko presentation includes multiple viewing and monitoring scenarios, including live monitoring at the visitor center, management-side monitoring through Patrol Server, and public visitor viewing through YouTube.

Different users require different viewing methods. Public visitors may prefer simple access through YouTube or a tourism website. Site managers may need a more operational interface for monitoring multiple locations, reviewing live conditions, or coordinating field response.

This flexibility is one of the key advantages of 360° live streaming in smart tourism. The same panoramic visual source can support public communication, visitor engagement, and management-side awareness, depending on how the video stream is integrated into the platform.

A traditional fixed-v iew camera usually shows only one direction, such as a bridge entrance, road section, scenic platform, or visitor facility. This can be useful, but it may not provide enough surrounding context for outdoor environments.

A 360° panoramic camera captures the surrounding scene from a single viewpoint. This allows visitors and managers to see not only one target area, but also the broader environment around it.

For scenic destinations, this can support:

Real-time scenic previews before arrival
Wider visibility around viewpoints and public areas
Weather, visibility, and crowd awareness
Remote observation of roads, entrances, and visitor facilities
More immersive digital tourism experiences
Live content for public communication and destination marketing
A visual foundation for future AI monitoring and safety applications

In the current public-facing Taroko website, visitors can select a scenic location and open the corresponding live view through YouTube. This is suitable for public access because YouTube provides a familiar viewing experience and makes live video easy to share.

For management-side applications, the same 360° live video architecture can also be extended into more operational workflows. Cupola360⁺ Patrol can support multi-camera viewing, panoramic monitoring, map or site-based navigation, event overlays, and embedded operational dashboards when a project requires a more controlled monitoring interface.

This distinction is important. YouTube is suitable for public viewing and outreach, while Patrol is better suited for management-side monitoring, internal operations, and system integration.

For smart tourism projects, both models can work together. A scenic destination may use YouTube for public-facing live views, while using Patrol or a similar management platform for operational monitoring, AI alerts, or internal coordination.

Although this case focuses on 360° live panorama and smart tourism, DP Smart’s Taroko platform concept also demonstrates how panoramic live imagery can be extended into AI and safety applications.

The presentation outlines AI smart monitoring features such as crowd counting, virtual fence alerts, and movement tracking. These functions can help managers understand visitor distribution, define restricted areas, and analyse movement patterns across panoramic scenes.

For large outdoor destinations, 360° visual data can also support disaster monitoring and visitor safety. DP Smart’s presentation highlights potential applications such as landslide detection, flood and rain monitoring, rescue coordination, and traffic control for high-mountain gorge terrain.

Beyond safety, long-term panoramic records may also contribute to sustainability communication, ecological monitoring, and ESG-related reporting. The presentation connects 360° panorama with conservation visibility, vegetation recovery records, biodiversity observation, and sustainability disclosure.

Key Highlights

Built a 720 live image platform for Taroko smart tourism
Used Cupola360 panoramic camera series for 360° live panorama
Supported 4K panoramic live views and outdoor-ready deployment
Streamed live views from eight key scenic locations across Taroko
Enabled public access through location-based viewing and YouTube Live
Supported visitor previews, weather and scenery awareness, and digital tourism engagement
Created a visual foundation for management-side monitoring, AI applications, disaster response, and sustainability use cases
Demonstrated how 360° live streaming can become part of a broader smart tourism platform

Note : This video shows Taroko 720 live image website, where visitors can select scenic locations and open the corresponding 360° live views. Click to experience the platform.

360° 全景影像座標映射：支援局部視角與 AI 偵測的空間定位技術

Mon, 20 Jul 2026 05:52:21 GMT

360° 全景相機能提供大範圍可視性，但有效的監控不只是取得完整畫面。局部視角、AI 偵測結果、註解與事件標記，都必須能夠與原始全景影像座標系統保持連結。

座標映射提供了這層連結。透過將局部視角座標轉換為 3D 方向，再投影回全景影像，系統能在局部檢視、AI 偵測與完整場景監控之間保留一致的空間脈絡。

對 AI 監控與遠端管理應用而言，360° 視覺不只是更寬的視角，而是能作為一層空間參照，幫助操作人員理解事件發生的位置、事件與周圍環境的關係，以及如何在更完整的脈絡下進行回應。

360° 全景相機能夠從單一視點提供完整場域影像。對監控應用而言，這種廣視角影像的價值在於能讓操作人員以較少的相機數量觀察更大的環境範圍，並降低傳統固定視角容易產生的盲區。

然而在實際監控情境中，使用者並不一定只觀看完整的全景畫面。操作人員可能會開啟某個局部視角，檢查特定設備、出入口或作業區域；AI 模型也可能針對局部裁切畫面進行物件偵測；使用者也可能在特定視角中加入標記、註解或事件點。

若這些局部資訊要在 360° 監控系統中真正發揮作用，就必須能夠被映射回原始的全景影像座標系統。如此一來，系統才能在完整全景畫面與局部視角之間維持一致的空間參照。

技術挑戰：如何連接局部視角與完整全景影像

一般固定式相機的影像具有固定畫面範圍。當某個物件在畫面中的特定像素位置被偵測到時，該位置就直接對應到同一張影像平面中的座標。

360° 全景影像則不同。全景影像是將周圍的球面場景投影到平面影像上，常見格式為等距柱狀投影（equirectangular projection）。在這種格式中，水平軸通常對應相機周圍的觀看方向，垂直軸則對應仰角與俯角。

當系統從 360° 影像中產生一個局部視角時，實際上是從全景球面中選取某個觀看方向，並轉換成類似一般相機畫面的透視視角。這個局部視角會有自己的中心方向、視野範圍、縮放比例與影像座標系統。

因此，系統必須解決一個座標映射問題：當使用者在局部視角中選取某個點，或 AI 模型在局部視角中偵測到物件時，系統需要計算該點或該偵測結果在原始 360° 全景影像中的對應位置。

如果缺少這層映射，局部視角會與完整場景脫節；若映射準確，局部檢視、AI 偵測與全景監控就能在同一套空間參照中運作。

什麼是局部視角映射?

局部視角映射，是指將局部透視畫面中的座標，轉換回完整 360° 全景影像座標的過程。

在監控系統中，這項技術可應用於多種資訊型態，例如：

使用者在局部視角中點選的位置
AI 偵測模型產生的 bounding box
人工加入的註解或事件標記
使用者選取的感興趣區域
事件回放或檢視時所對應的歷史位置

其核心目的，是保留「局部細節」與「完整場景」之間的空間關係。舉例來說，若 AI 模型在局部視角中偵測到人員，系統應能將該偵測結果顯示在全景影像中的正確位置；若操作人員在放大視角中標記某台設備或某個門口，該標記也應能對應回完整 360° 場景中的實際方位。這對 AI 偵測、事件回溯、遠端監控與多視角導覽都非常重要。

360° 全景影像的座標系統

局部視角的座標系統

常見的 360° 影像格式為等距柱狀投影。在這種格式中，完整的球面場景會被展開成一張矩形影像。

水平座標通常代表方位角，也就是相機周圍的 yaw 方向。完整的水平範圍對應 360°。垂直座標則代表俯仰角，也就是 pitch，從球面的上方延伸到下方。

簡單來說：

全景影像中的水平移動，對應向左或向右觀看相機周圍的環境。
全景影像中的垂直移動，對應向上或向下觀看。
全景影像中的每一個像素，都可以對應到 3D 空間中的一個方向。

也因為這樣的結構，局部視角與完整全景影像之間的對應，不能只用一般 2D 裁切的方式處理。系統必須考慮球面投影、觀看方向、視野範圍與透視轉換等因素。

局部視角是從全景影像中產生的透視畫面。它的視覺效果更接近一般相機畫面，只顯示 360° 場景中的特定方向與範圍。

一個局部視角通常會包含以下參數：

Center yaw：局部視角面向的水平方向
Center pitch：局部視角面向的垂直方向
Field of view：局部視角的視野寬度
Output width and height：局部視角輸出影像的尺寸
Pixel coordinates：某個點在局部視角影像中的位置

當某個點出現在局部視角中時，它的像素座標只在該局部畫面中具有意義。若要將該點還原到全景影像中，系統必須先將該像素位置解讀為一條從相機中心延伸出去的視線方向，再將這個方向投影回 360° 全景影像。

座標映射流程

局部視角映射可以理解為一連串座標轉換。實際實作方式會依相機型號、投影方式、鏡頭校正與軟體架構而有所不同，但整體邏輯大致相同。

將局部像素座標轉換為標準化視角座標
第一步，是將局部視角中的某個點，從像素座標轉換為相對於局部畫面中心的標準化座標。例如，若局部視角具有固定寬度與高度，畫面中的某個點可以被表示為相對於中心點的水平與垂直偏移量。這個標準化位置能讓系統判斷該點位於透視畫面中的哪個位置。

靠近局部畫面中心的點，通常對應局部視角的中心方向；靠近邊緣的點，則會對應到偏離中心方向較遠的視線方向，偏移程度會受到 field of view 影響。

將標準化視角座標轉換為 3D 方向向量
完成標準化後，系統會將該點轉換為 3D 空間中的方向向量。這個步驟會使用局部視角的 field of view 與透視投影參數。原本的 2D 座標會被視為一條從相機中心射向 3D 場景的視線。

接著，系統會根據局部視角的 center yaw 與 center pitch，將該方向旋轉到全景相機的座標系統中。此時，系統處理的已不只是影像中的像素位置，而是 360° 場景中的一個空間方向。

將 3D 方向向量轉換為全景影像座標
當 3D 方向向量確定後，系統便可將該方向轉換回球面角度，通常包含 yaw 與 pitch。這些角度接著會被映射到等距柱狀全景影像中的對應像素位置。最終結果，就是局部視角中的點、標註或 AI 偵測結果，在原始 360° 全景影像中的位置。

若要處理的是 bounding box，系統可能會依需求映射其中心點、四個角點，或沿著邊界取樣多個點，以提高偵測區域在全景影像中的呈現精度。

將 AI 偵測結果映射回完整場景

AI 偵測模型通常會處理一般透視影像或局部裁切區域。這樣做的原因是，多數 AI 模型原本就是以一般影像格式訓練，在局部視角中進行分析，往往比直接處理變形的全景影像更容易取得穩定結果。

然而，如果偵測結果只停留在局部視角中，操作人員可能會失去完整場景脈絡。系統必須知道該偵測結果在全景影像中的位置，才能讓事件判讀更具空間意義。

例如：

在管制區偵測到人員時，應能顯示其在完整 360° 場景中的正確位置。
PPE 偵測結果應能對應到實際發生的作業區域。
車輛或物件若出現在場域邊界附近，應能被放回更大的場域脈絡中檢視。
由局部 AI 模型觸發的警報，應能在全景影像與回放時間軸中被檢視。

透過座標映射，AI 偵測結果可以被連接回全景影像的空間參照，協助操作人員確認警報、理解周圍環境，並以更完整的脈絡進行事件回溯。

結語

Bounding Box 與偵測區域的處理

將單一點從局部視角映射回全景影像，相對來說較為直接。但若要處理 bounding box 或偵測區域，則需要更多考量。

在局部透視畫面中，bounding box 通常是一個矩形。然而，當它被投影回 360° 全景影像後，對應區域未必仍然是完美矩形，因為全景影像採用的是球面投影。

依照實際應用需求，系統可能採用不同處理方式：

映射 bounding box 的中心點，用來表示物件位置
映射 bounding box 的四個角點
沿著 bounding box 邊界取樣多個點
將偵測區域轉換為全景影像上的多邊形
以中心點作為事件標記，同時保留原始局部偵測畫面供視覺確認

合適的方法取決於應用所需的精度、介面設計與 AI 使用情境。若只是事件告警，中心點標記可能已足夠；若需要更精準的視覺疊圖或事件回放，則可能需要更完整的區域映射。

為什麼座標映射對監控應用很重要

標映射不只是數學轉換。在監控系統中，它會直接影響操作人員理解事件與回應事件的方式。

準確的座標映射可支援：

在完整全景影像中定位事件
驗證 AI 偵測結果
維持標註與事件標記的位置一致性
在局部視角與完整場景之間切換導覽
以空間脈絡進行事件回放
減少多個相似局部視角造成的判讀混淆
強化 AI 結果與視覺監控流程之間的整合

對遠端操作人員而言，這一點尤其關鍵。偵測結果不應只告訴使用者「偵測到了什麼」，也應該讓使用者知道「它發生在場域中的哪裡」。

在監控應用中，當局部視角、AI 偵測結果與事件標記都需要維持與完整 360° 場景的連結時，座標映射就會變得特別重要。

在智慧工廠與工業場域中，座標映射可協助操作人員將局部偵測或標註連接到產線區域、出入口、設備區、倉儲區與安全區域。在資料中心與設備機房中，局部視角可能聚焦於機櫃、走道、設備或門禁點，而全景影像則能保留整體位置脈絡。

對於工地、校園、公共空間、遠端設施與機器人巡檢系統，同樣的映射邏輯也能支援事件定位、事故回溯與更直覺的遠端監控流程。

Cupola360 的全景視覺平台，正是為了支援完整場景可視化與局部視角互動之間的連結而設計。透過維持局部視角與原始全景影像之間的座標關係，Cupola360 能將 360° 影像轉化為 AI 監控與遠端巡檢工作流程中的空間參照層。

這使得 AI 偵測、標註與事件標記不只是孤立的結果，而能被放回更完整的場域脈絡中檢視。

支援不同場域的 AI 監控應用

360° Panoramic Image Coordinate Mapping for Local View and AI Detection

Wed, 08 Jul 2026 02:31:09 GMT

For monitoring applications, coordinate mapping becomes especially useful when local views, AI detection results, and event markers need to remain connected to the full 360° scene.

In smart factories and industrial sites, this can help operators connect local detections or annotations with production areas, entrances, equipment zones, storage areas, and safety regions. In data centers and equipment rooms, local views may focus on racks, aisles, devices, or access points, while the panoramic image helps preserve overall location context. For construction sites, campuses, public spaces, remote facilities, and robotic patrol systems, the same mapping logic can support event localization, incident review, and more intuitive remote monitoring.

Cupola360’s panoramic vision platform is designed to support this connection between full-scene visibility and local-view interaction. By maintaining the relationship between local views and the original panoramic image, Cupola360 helps turn 360° video into a spatial reference layer for AI-enabled monitoring and remote inspection workflows.

This allows AI detections, annotations, and event markers to be reviewed not only as isolated results, but as part of the wider site context.

360° panoramic cameras provide full-scene visibility from a single camera point. For monitoring applications, this wide field of view is valuable because it allows operators to observe a larger environment with fewer blind spots.

However, in practical use, monitoring does not always happen directly on the full panoramic image. Operators may open a local view to inspect a specific area. AI models may process a cropped region for object detection. Users may add markers, annotations, or event points inside a selected view.

For these functions to be useful in a 360° monitoring system, the local information must be mapped back to the original panoramic image coordinate system. This allows the system to maintain a consistent spatial reference between the full panoramic scene and its local views.

The Challenge: Connecting Local Views with the Full Panoramic Image

A standard camera image has a fixed frame. If an object is detected at a certain pixel position, that position belongs directly to the same image plane.

A 360° panoramic image is different. The image represents a surrounding spherical scene projected onto a flat surface, often in an equirectangular format. In this format, the horizontal axis corresponds to the viewing direction around the camera, while the vertical axis corresponds to elevation.

When a local view is generated from a 360° image, it is essentially a perspective view looking toward a specific direction inside the panoramic sphere. This local view has its own center direction, field of view, zoom level, and image coordinate system.

This creates a coordinate-mapping problem. If a user selects a point inside the local view, or if an AI model detects an object inside that local view, the system needs to calculate where that point or detection belongs in the original 360° panoramic image.

Without this mapping, local views become disconnected from the full scene. With accurate mapping, local inspection, AI detection, and panoramic awareness can work within the same spatial reference.

What Local View Mapping Means

Local view mapping is the process of converting coordinates from a local perspective view back to the full 360° panoramic image.

In a monitoring system, this can apply to several types of information:

A user-clicked point inside a local view
A bounding box generated by an AI detection model
A manually added annotation or event marker
A selected region of interest
A replay or review point linked to a previous event

The purpose is to preserve the relationship between local detail and full-scene context. For example, if an AI model detects a person in a local view, the system should be able to display that detection at the correct position in the panoramic image. If an operator marks a specific machine or doorway in a zoomed view, that marker should still correspond to the correct location in the full 360° scene.

This is especially important for AI detection, event review, remote monitoring, and multi-view navigation.

Panoramic Image Coordinate System

Local View Coordinate System

A common 360° image format is the equirectangular projection. In this format, the entire spherical scene is unwrapped into a rectangular image.

The horizontal coordinate usually represents the azimuth angle, or yaw, around the camera. A full horizontal range corresponds to 360°. The vertical coordinate represents the elevation angle, or pitch, from the top to the bottom of the sphere.

In simplified terms:

Horizontal movement in the panoramic image corresponds to looking left or right around the camera.
Vertical movement corresponds to looking up or down.
Each pixel in the panoramic image can be related to a direction in 3D space.

Because of this structure, mapping between local views and the full panoramic image requires more than a simple 2D crop calculation. The system must account for spherical projection, viewing direction, field of view, and perspective transformation.

A local view is a perspective view generated from the panoramic image. It behaves more like a normal camera view, showing only a selected direction and field of view from the 360° scene.

A local view usually includes parameters such as:

Center yaw: the horizontal direction the local view is facing
Center pitch: the vertical direction the local view is facing
Field of view: how wide or narrow the local view is
Output width and height: the local view image size
Pixel coordinates: the position of a point inside the local view

When a point appears inside the local view, its pixel coordinate is only meaningful within that local image. To restore it to the panoramic image, the system must first interpret that pixel as a viewing ray, then project the ray back to the 360° image.

Coordinate Mapping Workflow

The mapping process can be understood as a sequence of transformations. The exact implementation may vary depending on camera model, projection method, lens calibration, and software architecture, but the general logic is similar.

Convert Local Pixel Coordinates to Normalized View Coordinates
The first step is to convert a point from local pixel coordinates into normalized coordinates relative to the center of the local view.

For example, if the local view has a width and height, a point inside that view can be represented by its horizontal and vertical offset from the center. This normalized position tells the system where the point lies within the perspective view.

A point near the center of the local image corresponds to the local view’s center direction. A point near the edge corresponds to a direction farther away from the center, depending on the field of view.

Convert Normalized View Coordinates to a 3D Direction Vector
After the local point is normalized, the system converts it into a 3D direction vector. This step uses the local view’s field of view and perspective projection parameters. The normalized 2D point is treated as a ray extending from the camera center into the 3D scene.

The local direction is then rotated according to the local view’s center yaw and pitch. This produces a direction vector in the panoramic camera’s coordinate system. At this stage, the system is no longer only working with a pixel position. It is working with a spatial direction inside the 360° scene.

Convert the 3D Direction Vector to Panoramic Coordinates
Once the 3D direction vector is known, the system can convert it back into spherical angles, usually yaw and pitch. These angles are then mapped to the corresponding pixel coordinates in the equirectangular panoramic image.

The result is a position in the original 360° image that corresponds to the selected point, annotation, or AI detection in the local view.

For a bounding box, this process may be applied to key points such as corners or sampled boundary points, depending on how accurately the detection area needs to be represented on the panoramic image.

AI detection models often operate on standard perspective images or cropped regions. This is useful because many AI models are trained on conventional image formats and may perform better when analysing a local view instead of a distorted panoramic image.

However, if the detection result remains only inside the local view, operators may lose the full-scene context. The system needs to know where that detection belongs in the panoramic image.

For example:

A person detected in a restricted zone should be shown at the correct location in the full 360° scene.
A PPE detection result should remain linked to the work area where it occurred.
A vehicle or object detected near a facility boundary should be placed back into the wider site context.
An alert generated from a local AI model should be reviewable in the panoramic image and playback timeline.

Coordinate mapping allows AI detection results to be connected back to the panoramic spatial reference. This helps operators verify alerts, understand surrounding conditions, and review incidents with better context.

Supporting AI-Enabled Monitoring Across Different Sites

Mapping AI Detection Results Back to the Full Scene

Mapping a single point from a local view to the panoramic image is relatively straightforward. Mapping a bounding box or detection region requires additional consideration.

A bounding box in a local perspective view is rectangular. After projection back to the panoramic image, the corresponding region may not remain a perfect rectangle because the panoramic image uses spherical projection.

Depending on the application, the system may choose different methods:

Map the center point of the bounding box to indicate the object location
Map all four corners of the bounding box
Sample multiple points along the bounding box boundary
Convert the region into a polygon on the panoramic image
Use the mapped center point for event markers while keeping the original local detection view for visual confirmation

The appropriate method depends on the required precision, interface design, and AI application. For event alerting, a center marker may be sufficient. For visual overlay or detailed review, a more accurate projected region may be needed.

Coordinate mapping is not only a mathematical conversion. In monitoring systems, it directly affects how operators understand and respond to events.

Accurate mapping can support:

Event localization in the full panoramic image
AI detection verification
Consistent annotation and marker placement
Local view and full-scene navigation
Incident replay with spatial context
Reduced confusion between similar local views
Better integration between AI results and visual monitoring workflows

For remote operators, this is especially important. A detection result should not only say what was detected. It should also show where it happened in relation to the surrounding environment.

Handling Bounding Boxes and Detection Regions

Why Coordinate Mapping Matters for Monitoring Applications

Conclusion

360° panoramic cameras provide wide-area visibility, but effective monitoring requires more than a full image. Local views, AI detection results, annotations, and event markers need to remain connected to the original panoramic coordinate system.

Coordinate mapping provides this connection. By converting local view coordinates into 3D directions and projecting them back to the panoramic image, the system can preserve spatial context across local inspection, AI detection, and full-scene monitoring.

For AI-enabled and remote monitoring applications, this makes 360° vision more than a wide-angle view. It becomes a spatial reference layer that helps operators understand where events happen, how they relate to the surrounding scene, and how to respond with better context.

Japan’s Growing Bear Encounters: How 360° AI Vision Can Support Safer Field Awareness

Mon, 06 Jul 2026 05:56:30 GMT

Using panoramic cameras and AI detection to monitor outdoor areas, reduce blind spots, and respond with better context.

In Japan, bear encounters have become a growing safety concern, not only in mountain areas but also around residential communities, factories, tourism sites, and remote facilities. As wildlife activity moves closer to human environments, early detection and wider field awareness are becoming increasingly important.

For operators managing outdoor sites, the challenge is not only detecting a bear after it appears. The real challenge is understanding the surrounding area: where the risk is coming from, how close it is to people or buildings, and how the situation is changing over time.

At a recent exhibition in Osaka, Cupola360’s Japan partner iiTEC Co., Ltd. showcased an application using Cupola360 360° cameras with AI to support real-time bear detection and field monitoring. This example shows how panoramic vision can support a broader category of wildlife safety and outdoor monitoring applications.

Wildlife safety is becoming an important issue for many outdoor and semi-remote environments. Residential areas near mountains, factories located close to natural surroundings, tourism sites, parks, campsites, and public facilities may all face unexpected wildlife activity.

In these environments, safety teams often need to monitor wide and open areas. A risk may appear from any direction, move quickly, or enter from a place that is not covered by a traditional camera angle.

For bear detection, this is especially important. Early awareness can give operators more time to notify staff, guide people away from the area, and coordinate a safer response.

The Challenge: Bear Encounters Are Moving Closer to Human Environments

Why Fixed-View Cameras May Leave Outdoor Blind Spots

Traditional cameras are useful for focused monitoring, but they usually cover only one direction at a time. In outdoor environments, this can create blind spots between camera angles or behind objects, trees, buildings, vehicles, and uneven terrain.

To cover a wider site, teams may need to install multiple cameras and monitor several separate video feeds. Even then, operators may still need to switch between views to understand the full situation.

For field safety, fragmented visibility can slow down response. When a potential risk appears, operators need to understand the full scene quickly, not only a single fixed view..

The Application: 360° Cameras with AI Bear Detection

A 360° panoramic camera provides a complete surrounding view from a single point. When combined with AI detection, the system can continuously analyse a wider scene and help identify potential risks earlier.

In a bear detection scenario, AI can be configured to detect specific objects or movements within defined areas. When a potential risk is identified, the system can support real-time alerts for field operators.

This type of application can help support:

Bear detection in defined outdoor areas
Real-time alerting for field operators
Full-scene video records for event review
Wider site awareness with fewer blind spots
Remote monitoring for facilities, tourism sites, and public spaces

Note: Conceptual illustration for demonstration purposes only; not an actual deployment image or verified detection result.

The value of 360° vision is not only wider coverage. It also helps operators understand the full context around an event.

In outdoor environments, context matters. Operators may need to know where an animal entered, whether people are nearby, how close the risk is to buildings or work areas, and whether the situation is moving toward or away from a critical zone.

A panoramic view helps provide this context in one continuous scene. Instead of switching between multiple disconnected camera angles, operators can view the surrounding environment more intuitively and make faster decisions.

This is especially useful for sites where staff cannot constantly monitor every direction in person, such as remote facilities, mountain areas, factories near natural surroundings, parks, tourism destinations, and public spaces.

Beyond Bear Detection: Wider Outdoor Safety Applications

Although this example focuses on bear detection in Japan, the same 360° AI vision architecture can support many other outdoor safety applications.

With the right AI model and integration, panoramic visual data can be used for:

Wildlife monitoring
Intrusion detection
Perimeter security
Factory boundary monitoring
Warehouse and logistics yard monitoring
Smart city public safety
Tourism site monitoring
Remote facility awareness

Many of these applications share the same core challenge: teams need earlier detection, fewer blind spots, and better situational awareness before they can respond effectively.

Related Resources

Cupola360’s Role: A 360° Vision Platform for AI Partners

In partner-driven AI applications, Cupola360 provides the 360° panoramic camera and visual platform that AI partners, system integrators, and solution providers can build on top of.

A 360° camera is more than a camera. It can serve as the visual foundation for AI applications by providing wide-area panoramic data for detection, alerting, monitoring, and event review.

For wildlife detection and other outdoor safety applications, this means partners can combine their AI capabilities with Cupola360’s panoramic vision to create solutions for specific industries, sites, and operating environments.

Cupola360 is excited to see partners such as iiTEC bringing innovative 360° AI vision applications into the field. As outdoor safety challenges continue to evolve, panoramic vision can help create more flexible and context-aware monitoring systems across different industries.

Why Panoramic Vision Matters for Field Awareness

Kaohsiung Traffic Monitoring : 360° Cameras for Map-Based City Awareness

Fri, 03 Jul 2026 04:18:21 GMT

How panoramic cameras can complement traditional traffic views and third-party city platforms.

Kaohsiung, Taiwan is one of the country’s major southern cities, with busy intersections, public transportation routes, tourism areas, parking zones, and dense urban traffic activity.

To support real-time traffic awareness, Kaohsiung City provides a map-based traffic information platform that brings together road status, traffic events, parking information, vehicle detector data, traditional camera feeds, and 360° panoramic camera views.

In this application reference, Cupola360 shares how 360° panoramic cameras can work as one visual layer within this type of map-based traffic platform. The goal is not to replace traditional cameras or existing city systems, but to complement them with wider scene awareness.

Traditional traffic cameras are useful for focused views of specific roads, lanes, or directions. 360° panoramic cameras add broader context by capturing the full surrounding environment from a single camera point. When used together, they help users move from checking isolated camera angles to understanding the full scene around an intersection or public space.

The Challenge: Traffic Monitoring Needs Both Detail and Context

Urban traffic environments are complex. A single intersection may include vehicles moving from multiple directions, scooters, buses, pedestrians, parking activity, nearby public facilities, and unexpected incidents.

Traditional cameras provide important detail, but one camera angle may not explain the whole situation. Congestion may be related to a nearby event, an incident on another road, a turning lane, pedestrian movement, or surrounding traffic flow.

For city monitoring, operators and users need both focused views and wider context. This is where 360° panoramic vision can add value.

The Application: 360° Views Inside a Map-Based Traffic Platform

Note: Screenshot adapted and translated from Kaohsiung City Real-Time Traffic Information Platform. Translation is provided for illustrative purposes only. Original platform: https://traffic.tbkc.gov.tw/

Why 360° Panoramic Cameras Matter

Supporting Existing Infrastructure, Not Replacing It

Note

Key Highlights

Related Resources

For smart city and traffic monitoring applications, adding more cameras is not always the best answer. What matters is whether the visual information helps users understand the situation faster and more clearly.

360° panoramic cameras can support:

Wider intersection awareness from one camera point
Better context around traffic flow and roadside activity
Reduced blind spots in complex public environments
Faster event verification when incidents occur nearby
Stronger integration with existing cameras and map-based systems
More flexible monitoring for intersections, tourism areas, parking lots, and public spaces

The value of panoramic vision is not only in seeing a wider image. It is in helping people understand the relationship between the camera view, the location, and the surrounding situation.

This use case shows an important direction for smart city monitoring: 360° cameras do not need to replace traditional cameras. Instead, they can work alongside existing infrastructure and third-party platforms.

Traditional cameras provide detail.

360° panoramic cameras provide context.

Map-based platforms connect the visual information with location, traffic status, and event data.

Together, these elements can create a more complete monitoring experience for traffic management, public space awareness, and remote city operations.

For Cupola360, this reference demonstrates how panoramic camera technology can be integrated into broader monitoring ecosystems to help teams and users move from separate camera feeds to fuller site awareness.

Cupola360 provides 360° panoramic camera technology for visual integration scenarios. The overall platform, map data, traffic layers, and system operation are managed by the respective platform owner or service provider.

360° panoramic cameras used as a visual layer within a map-based traffic platform
Traditional camera views and 360° views support different monitoring needs
Road traffic status, event markers, parking information, and detector data provide additional context
Map-based access helps users find camera points by location
Panoramic vision helps users understand full-scene conditions around intersections and public areas
Demonstrates how Cupola360 cameras can complement existing traffic and city monitoring infrastructure

On Kaohsiung’s real-time traffic platform, users can access different map layers, including traffic status, traditional roadside camera views, 360° intersection camera points, parking information, traffic incidents, vehicle detector information, and other event markers.

This map-based interface makes the experience more intuitive. Users can start from the location they care about, check road conditions through color-coded traffic status, and open relevant camera views when they need more visual context.

In this setup, 360° panoramic cameras help extend the platform’s visual coverage. Instead of showing only one fixed direction, a 360° camera can provide a fuller view of the intersection or surrounding public area.

YCM × Cupola360: Panoramic Smart Inspection for Safer Manufacturing Operations

Fri, 03 Jul 2026 04:00:11 GMT

Connecting factory visibility, safety detection, and work reporting through panoramic smart inspection.

YCM, also known as Yeong Chin Machinery, is a Taiwan-based machine tool manufacturer. To support smart manufacturing and workplace safety management, YCM introduced a panoramic smart inspection system with Cupola360 and DotZero.

In this deployment, Cupola360 provides the panoramic visual layer, while DotZero supports smart manufacturing integration and AI-enabled workflow applications. Together, the solution helps YCM improve real-time production visibility, safety awareness, and digital factory operations.

The Challenge: Connecting Production Visibility with Safety Management

In manufacturing environments, production management and workplace safety are often handled through separate systems. Factory teams may need to monitor production status, check whether workers follow safety rules, respond to abnormal situations, and record work progress across different tools or manual processes.

When the site is large and workflows are complex, fixed camera views and manual reporting may not provide enough visibility or speed. Teams need a more connected way to see the site, understand what is happening, and respond quickly when safety or production issues occur.

The Solution: Panoramic Smart Inspection with Partner Integration

Cupola360 RX1000F panoramic cameras help helps provide wider site visibility across the factory environment. Instead of relying only on narrow camera angles, teams can view production areas with more complete visual context and better understand how people, equipment, and workflows interact on site.

With DotZero’s smart manufacturing integration, the panoramic visual layer can be connected with AI-enabled applications and factory workflows. This allows the system to support real-time production tracking, safety recognition, event alerts, and work reporting in a more integrated way.

Real-Time Production Visibility

PPE Recognition and Safety Alerts

Gesture-Based Work Reporting

Supporting Smart Manufacturing Transformation

Key Highlights

Related Resources

Through real-time panoramic imaging, factory teams can view operation areas from a more immersive perspective. This helps improve site transparency and gives managers a clearer way to understand production flow, worker movement, and on-site conditions.

For manufacturing teams, this means the system is not only used for video monitoring. It also supports operational awareness, helping teams review factory activity with broader context and respond with more confidence.

The system supports AI-assisted safety applications, including PPE recognition. It can help identify whether workers are properly wearing required safety equipment such as helmets and safety vests in designated areas.

When abnormal behavior or unsafe conditions are detected, the system can trigger alerts to help teams respond faster. This strengthens workplace safety management and reduces reliance on manual observation alone.

Another key application is gesture-based work reporting. Through gesture recognition, workers can complete reporting actions using natural movements, making the reporting process more convenient and reducing the need for physical contact with devices.

This is especially useful in factory environments where workers may be moving between stations, handling materials, or operating under safety and hygiene requirements. By making reporting easier, the system helps connect frontline operations with digital management workflows.

For YCM, panoramic smart inspection demonstrates how real-time visual context, AI recognition, safety alerts, and work reporting can work together in factory operations.

For Cupola360 and DotZero, this use case shows how panoramic vision can integrate with smart manufacturing platforms to support more flexible industrial applications. The value is not only in monitoring the site, but in helping manufacturers connect visibility, safety, and workflow data as part of their digital transformation journey.

Deployed 14 Cupola360 RX1000F panoramic cameras across key factory areas
Combined Cupola360 panoramic vision with DotZero’s smart manufacturing integration
Supported real-time production visibility and site transparency
Enabled AI-assisted PPE recognition for helmets and safety vests
Supported safety alerts for abnormal behavior or risky situations
Enabled gesture-based work reporting for more convenient factory operations
Helped strengthen smart manufacturing and workplace safety management

Taiwan Galvanizing × Cupola360: Panoramic Vision for AI-Enabled Factory Safety

Thu, 02 Jul 2026 06:11:29 GMT

Improving safety compliance and process visibility across high-risk galvanizing operations.

Taiwan Galvanizing Technology Co., Ltd. operates in a demanding industrial environment where worker safety, process visibility, and production control are all critical. Hot-dip galvanizing involves heavy materials, lifting operations, process tanks, temperature control, and strict safety requirements across the factory floor.

To improve site visibility and support safer operations, Taiwan Galvanizing deployed 16 Cupola360 RX1000P panoramic cameras across key factory areas. The system works with Synology NAS and third-party AI detection to support wide-area monitoring, safety recognition, and process awareness.

The Challenge: Safety and Visibility in High-Risk Manufacturing

In hot-dip galvanizing and metal processing environments, safety risks can come from multiple directions at the same time. Workers move through production areas, materials are lifted and transported, and process tanks must be monitored under specific operating conditions.

Traditional fixed-view cameras can capture selected angles, but they may not provide enough context across large production spaces, entrances, exits, crane operation areas, and process zones. When teams need to review an event or confirm whether safety rules are being followed, fragmented camera views can make the process slower and less complete.

For this type of industrial site, monitoring needs to go beyond passive recording. The factory needs a way to improve visibility, support safety compliance, and connect visual information with production conditions.

The Solution: RX1000P Panoramic Cameras with NAS Integration

Cupola360 RX1000P panoramic cameras were installed across the factory to monitor important areas such as entrances, exits, production zones, and process-related spaces. With 360° panoramic coverage, each camera can provide wider visibility and help reduce blind spots compared with traditional fixed-angle views.

The cameras are connected with Synology NAS for video storage and management, creating a practical foundation for factory monitoring and event review. This setup allows the factory to retain panoramic visual records while supporting integration with additional AI and operational systems.

AI-Enabled Safety Detection with Third-Party Systems

Crane, Material, and Process Area Monitoring

From Factory Monitoring to Digital Transformation

Note

Key Highlights

Related Resources

By integrating Cupola360 panoramic cameras with third-party AI detection, the factory can support safety-related recognition such as helmet and safety vest detection in designated areas.

Cupola360’s role is to provide the panoramic visual layer. The wider scene view helps give AI systems and operators more complete visual context, especially in areas where people, equipment, and production activities overlap.

When PPE requirements are not met, the integrated system can support alerts and follow-up actions, helping factory teams strengthen daily safety compliance.

The system also supports visual monitoring around crane and material handling areas. In production environments where suspended materials, lifting angles, and worker movement may overlap, 360° visibility helps operators better understand the relationship between equipment movement, people nearby, and production zones.

The panoramic visual layer can also be displayed together with process information from existing or third-party systems, such as tank conditions, concentration, iron ion levels, and temperature. This helps teams review visual context alongside production parameters and make more informed operational decisions.

For Taiwan Galvanizing, panoramic vision supports more than factory surveillance. It creates a foundation for safer production monitoring, AI-assisted safety checks, process visibility, and event review.

By combining panoramic video, NAS storage, third-party AI detection, and process information, the system helps teams monitor operations from production workflow to safety compliance. Over time, this type of visual and operational data can support improvements in process optimization, waste reduction, energy efficiency, and digital transformation.

Cupola360 provides the RX1000P panoramic camera and visual integration layer in this deployment. NAS storage is supported through Synology, and AI detection is provided by a third-party AI partner.

Deployed 16 Cupola360 RX1000P panoramic cameras
Integrated with Synology NAS for video storage and management
Worked with third-party AI detection for helmet and safety vest recognition
Monitored entrances, exits, and key production areas
Improved visibility around crane and material handling zones
Displayed panoramic views together with process-related data such as concentration, iron ion levels, and temperature
Supported factory safety, event review, and process visibility

Note : This diagram is a conceptual camera layout for reference only and does not represent the actual factory deployment plan.

Panoramic Vision for Cashier Operation Monitoring | Cupola360 Smart Retail Use Cases

Fri, 12 Jun 2026 09:22:42 GMT

Helping retail teams connect storewide visibility, AI detection, and cashier activity review.

For retail stores, the cashier area is one of the most important points of operation. It is where customer transactions happen, cash is handled, staff activity is recorded, and store teams often need clear evidence when discrepancies occur.

In this retail deployment, Cupola360 panoramic cameras were integrated with an AI partner’s detection interface to support cashier operation monitoring across store locations. The system combines a wide 360° store view with focused cashier-area monitoring, helping operators understand both the overall store environment and the details of specific transaction areas.

The Challenge: Cashier Activity Needs Better Context

Traditional retail surveillance often relies on fixed-view cameras near the cashier counter. These cameras can capture the checkout area, but they usually provide only a narrow view. When store teams need to review an event, they may have to switch between different camera feeds to understand what happened before, during, and after the transaction.

For cashier operations, context matters. Teams may need to know when the cash drawer was opened, how long it stayed open, how many people were present, and whether the action followed store policy. Without a connected view, reviewing these situations can be time-consuming and incomplete.

The Solution: Panoramic Store View with AI Cash Drawer Monitoring

Cupola360 provides a wide 360° panoramic view of the store environment, while the AI partner’s interface adds cashier-area analytics and event monitoring. In the system, operators can view the full store scene and click into specific cashier points to check more detailed views of the checkout area.

The AI monitoring logic is designed around cashier operation rules. For example, when the cash drawer is opened, the system can check whether the required number of staff members are present. If the cash drawer remains open beyond a defined time or the staffing condition is not met, the system can generate an alert.

The interface can also track drawer open and close activity, staff presence, and event counts, giving retail teams a clearer way to monitor cashier operations and review exceptions.

Supporting Accountability and Faster Review

Future Expansion for Smart Retail

Key Highlights

Related Resources

This use case helps retail teams improve cashier operation visibility without relying only on manual review. When a transaction discrepancy or operational question occurs, managers can use the system to review the timeline, check the relevant cashier area, and understand the surrounding store context.

The goal is not simply to record video. It is to help stores review events more efficiently, clarify responsibilities, and support fairer internal operation management.

With panoramic visibility, teams can also understand what happened around the cashier area, not just inside a single narrow camera frame. This makes event review more complete and helps reduce blind spots in store operations.

This deployment is currently planned for broader rollout across more retail locations. As the system expands, the same panoramic vision and AI integration approach can support more smart retail applications.

Beyond cashier monitoring, the platform can be extended to support loss prevention, suspicious behavior detection, customer flow analysis, staff safety, remote store inspection, and incident review. With 360° panoramic vision as the visual foundation, retail teams can build a more connected and scalable approach to store management.

For this retail use case, Cupola360 acts as the panoramic visual layer that helps connect storewide awareness, AI detection, cashier operation monitoring, and event review into one more practical workflow.

Integrated Cupola360 panoramic cameras with an AI partner’s detection interface
Supported 360° storewide visibility and focused cashier-area review
Monitored cash drawer opening activity and staff presence
Generated alerts based on cashier operation rules
Helped managers review transaction discrepancies with clearer context
Planned for broader rollout across additional retail locations
Can be extended to more smart retail applications such as loss prevention and remote inspection

Panoramic Vision for AI Server Lighthouse Factory Management | Cupola360 x Ingrasys

Fri, 12 Jun 2026 06:31:27 GMT

Turning factory visibility, AI, and production data into a smarter way to manage operations.

Ingrasys, part of Foxconn Technology Group, is a global provider of AI computing solutions and cloud infrastructure. At its AI Server Lighthouse Factory in Nanching, Ingrasys wanted to create a smarter way to monitor factory operations, connect production data, support AI analysis, and present the factory through a more immersive experience.

The Challenge: Fragmented Factory Visibility

In large manufacturing environments, traditional CCTV systems often require many fixed-view cameras to cover production lines, equipment areas, material flow, and key operation zones. Even with multiple cameras installed, managers may still need to switch between different video feeds to understand what is happening on site.

This fragmented view makes it harder to connect factory conditions with real-time production data. For smart manufacturing, teams need more than video recording. They need a clearer way to see the factory, understand the data, and manage operations remotely.

The Solution: 720° Panoramic Coverage with RX1000F

To improve visibility, Ingrasys deployed more than 50 Cupola360 RX1000F 720° panoramic cameras across the factory. With a unified panoramic spatial view, each camera can cover a wider area and provide more complete visual context.

Based on the Nanching factory deployment, Cupola360 helped reduce camera deployment by more than 60% and cut AI processing needs by over 50%. With fewer cameras and more complete visual input, the system supports both wider coverage and more efficient AI analysis.

Connecting Factory Data with Real-Time Dashboards

Supporting Immersive Factory Tours and Remote Management

AI Detection, Zone Alerts, and Event Review

An Open Platform for Smart Manufacturing

Key Highlights

Related Resources

Beyond visual monitoring, Cupola360 also connects factory images with machine and SFC data. Production information can be displayed through dynamic real-time dashboards, allowing teams to view factory status, machine data, and spatial context together.

This helps managers make faster, more informed decisions without separating production data from the actual factory scene. Operators can also use web iframe integration to view external systems and third-party web pages directly within the Cupola360 interface, creating a more unified platform for monitoring and remote management.

Cupola360 also enables a more immersive and interruption-free factory tour experience. Customers, partners, and visitors can explore production areas remotely without entering sensitive FAB or manufacturing zones.

Through panoramic views and multimedia content, Ingrasys can showcase its production strengths while keeping operations protected and uninterrupted. For internal teams, the same visual layer supports remote audits, site review, and factory management across different areas.

AI detection further supports daily factory management. With person detection, the system can monitor designated zones, send alerts when limits are exceeded, and notify teams again when the situation is resolved.

Operators can also move between cameras with intuitive directional controls and use an interactive mini map to understand the factory layout. For event review and AI validation, teams can replay 360° recordings with optional AI overlays to compare events, validate detection accuracy, and improve incident analysis.

Cupola360’s open platform gives Ingrasys flexibility to configure NVR and AI settings, enable or disable features, and integrate its own or third-party AI modules.

For Ingrasys, Cupola360 is not just a camera system. It acts as a panoramic visual layer that connects factory space, AI analysis, machine data, SFC information, dashboards, remote navigation, and immersive factory management into one more intuitive experience.

Deployed 50+ Cupola360 RX1000F panoramic cameras
Reduced camera deployment by more than 60%
Cut AI processing needs by over 50%
Integrated panoramic views with machine and SFC data
Created real-time dashboards for factory visualization
Supported web iframe integration with external systems
Enabled immersive factory tours without interrupting operations
Supported person detection, zone alerts, and 360° event replay

Cupola360 Connect 2026 : 首屆合作夥伴論壇於 COMPUTEX 期間圓滿落幕

Fri, 12 Jun 2026 04:02:45 GMT

串聯來自不同市場與產業的合作夥伴，一起探索 360° 全景 AI 視覺與 Reality Remote Management 的實際應用機會。

Cupola360 Connect 2026 已於 COMPUTEX 2026 期間圓滿落幕。對 Cupola360 來說，這是一個重要里程碑，因為這不只是一次活動，更是我們首次舉辦的合作夥伴論壇。

這次論壇邀集來自台灣、日本、馬來西亞與美國的合作夥伴與講者，也有講者從杜拜遠端參與。來自不同市場、不同產業背景的夥伴齊聚一堂，圍繞同一個核心議題交流：360° 全景 AI 視覺如何真正走進現場，支援更多實際應用。

從產品展示走向夥伴共創

對 Cupola360 而言，Cupola360 Connect 不只是簡報分享，也不只是單向介紹產品。這場論壇更重要的意義，是與合作夥伴一起討論如何把全景視覺、AI、軟體、系統整合與現場經驗結合起來，讓技術真正落地到不同產業場域。

在論壇中，我們分享了 Cupola360 對 Reality Remote Management 的願景，也與夥伴一起探討技術整合與場域應用的可能性。討論主題橫跨 AI vision、surveillance、robotic patrol、smart factory、smart livestock、data centers、immersive tourism 等不同應用方向。

這些主題看似分散，但背後其實指向同一個需求：越來越多產業需要更即時、更完整、更具現場脈絡的遠端場域管理能力。

為什麼合作夥伴生態系很重要

隨著遠端場域管理在各產業中的重要性提升，單一產品或單一技術很難獨立完成所有需求。真正能落地的解決方案，往往需要不同角色共同參與。

系統整合商理解客戶現場與部署條件。

AI 夥伴提供辨識、分析與自動化能力。

機器人與無人載具夥伴讓視覺系統能進入更難抵達的場域。

產業解決方案夥伴則能把技術轉化成更貼近場景的應用。

每一位合作夥伴都帶來不同拼圖，而 Cupola360 希望提供的是能串接這些拼圖的 360° 全景視覺基礎。透過更完整的現場影像、更直覺的空間脈絡，以及可延伸整合的軟硬體架構，我們期待與夥伴一起把新技術變成真正可部署、可營運、可擴展的解決方案。

從論壇出發，拓展更多真實場域應用

Cupola360 Connect 2026 的意義，不只在於完成第一場合作夥伴論壇，而是讓我們更清楚看見夥伴生態系的力量。

不同市場的夥伴帶來不同需求。

不同產業的應用場域帶來不同挑戰。

不同技術團隊則帶來不同整合方式與創新可能。

從智慧工廠到資料中心，從機器人巡檢到智慧畜牧，從沉浸式觀光到公共安全，360° 全景 AI 視覺正在走向更多真實環境。這些應用不只是展示技術，也代表未來遠端場域管理將更依賴跨領域合作。

感謝所有參與的夥伴與來賓

感謝所有講者、合作夥伴與來賓參與 Cupola360 Connect 2026。你們的想法、回饋與支持，讓我們的首屆合作夥伴論壇更具意義。

我們期待未來與全球更多合作夥伴創造新的機會，將 360° 全景 AI 視覺帶進更多真實場域，協助產業以更完整的視野理解現場、管理現場，並開創更多可落地的智慧應用。

Cupola360 亮相 COMPUTEX 2026：Omni Speed Dome 首次發表，與 ASPEED 聯合展出三大應用展示區

Fri, 12 Jun 2026 03:59:04 GMT

從 360° 全景視覺、AI、IoT 控制到機器人巡檢，Cupola360 於 COMPUTEX 2026 展示 Reality Remote Management 的最新產品與應用想像。

Cupola360 於 COMPUTEX 2026 與 ASPEED Technology 聯合展出，展示 360° 全景視覺、AI、IoT 控制與行動巡檢平台如何整合，支援更智慧的遠端場域管理。

今年對 Cupola360 來說特別重要，因為我們首次公開展示全新產品 Omni Speed Dome。這款次世代 360° 全景球型攝影機，將全景視野與重點區域追蹤結合，為需要即時掌握現場的應用場域，提供更快速、更完整的監控視角。

Omni Speed Dome 首次亮相

傳統機械式 Speed Dome 通常需要透過實體旋轉來切換視角，因此在轉向、追蹤或回看不同區域時，可能產生時間差，也容易錯過畫面之外的現場脈絡。

Omni Speed Dome 則以 360° 全景監控為核心設計，能在不依賴傳統旋轉等待的情況下，提供即時全景視野。這有助於操作人員減少監控盲區、維持完整 situational awareness，並在同一畫面中掌握多個重點區域。

透過 Omni Speed Dome，團隊可以先看見完整現場，再針對關鍵目標進行放大、鎖定與追蹤。對於工廠、公共空間、資料中心、機器人巡檢與遠端場域管理而言，這樣的視覺架構能同時兼顧「全局掌握」與「重點追蹤」。

三大 Demo Zones：從產品展示到應用場景

在 COMPUTEX 2026 展期間，Cupola360 也規劃了三大展示區，讓參觀者能從不同應用角度理解 360° 全景 AI 視覺的實際價值。原文列出三大展示區為 Reality Remote Management、360° Immersive Surveillance 與 AI Vision for Robotic Patrol。

現場遠端管理 (Reality Remote Management)

現場遠端管理 (Reality Remote Managemen)t 展示的是 Cupola360 對遠端場域管理的核心想像：不只是遠端看影像，而是將全景視覺、AI、即時資訊與遠端控制整合在一起，協助團隊掌握分散場域的真實狀況。

當管理者需要同時面對多個站點、不同設備、現場事件與營運資訊時，傳統監控畫面往往不夠直覺。RRM 的價值在於，讓使用者能在更完整的現場脈絡中理解事件，並更快做出判斷與應對。

360° 沉浸式監控

360° 沉浸式監控展示全景視覺如何支援大範圍場域監控。相較於分散的固定視角，全景影像能提供更完整的現場背景，讓操作人員不只看到單一角度，而是能理解事件周圍的空間關係。

這類應用特別適合工廠、展場、公共空間、校園、倉儲、資料中心與其他需要廣域監控的環境。

AI Vision for Robotic Patrol

AI Vision for Robotic Patrol 展示 Cupola360 相機與相機模組如何整合至機器人、無人載具與行動巡檢平台。透過行動平台搭配 360° 全景視覺，系統能在移動過程中取得更完整的環境視野，支援巡檢、遠端查看與 AI 視覺應用。

這對機器人巡檢、無人車、無人機、智慧工廠與高風險場域特別重要。當移動平台不只往前看，而是能持續掌握周圍環境，巡檢系統就能具備更完整的現場感知能力。

不只是展出產品，而是展示現場遠端管理的未來

對 Cupola360 而言，COMPUTEX 2026 不只是一次產品展示，更是與合作夥伴、客戶與產業專業人士交流現場遠端管理願景的重要機會。

從 Omni Speed Dome 到三大 Demo Zones，Cupola360 希望呈現的不只是單一相機或單一軟體功能，而是一套能連接現場、影像、AI、控制與行動平台的遠端場域管理架構。

感謝所有來到 ASPEED × Cupola360 聯合展位的夥伴與朋友。未來，我們也期待將 360° 全景 AI 視覺帶進更多真實應用場域，協助團隊更完整地看見現場、更快速地理解事件，並更有效地管理分散環境。

Cupola360 Connect 2026: Our First Partner Forum Successfully Wrapped Up During COMPUTEX

Thu, 11 Jun 2026 08:19:46 GMT

This is a subtitle for your new post

Cupola360 Connect 2026 successfully wrapped up during COMPUTEX 2026, marking an important milestone for us: our first partner forum.

We were grateful to bring together partners and speakers from Taiwan, Japan, Malaysia, and the United States, with one speaker even joining us remotely from Dubai. It was exciting to see people from different markets and industries come together around one shared topic: how panoramic AI vision can support real-world applications.

Throughout the forum, we explored Cupola360’s vision for Reality Remote Management , along with technology integration opportunities and field applications across AI vision, surveillance, robotic patrol, smart factory, smart livestock, data centers, immersive tourism, and more.

For us, Cupola360 Connect was not just about presentations. It was about building a stronger ecosystem with partners who understand the value of combining vision, AI, software, integration, and field experience.

As remote site management becomes more important across industries, we believe collaboration will play a key role in turning new technology into practical solutions. From system integrators and robotics partners to industry solution providers, every partner brings a different piece of the puzzle.

A big thank you to all speakers, partners, and guests who joined us. Your ideas, feedback, and support made our first partner forum meaningful.

We look forward to creating more opportunities with partners around the world and bringing 360° panoramic AI vision into more real-world environments.

Cupola360 at COMPUTEX 2026: New Product Launch, Joint Booth with ASPEED, and Live Demo Zones

Thu, 11 Jun 2026 08:16:45 GMT

This is a subtitle for your new post

Cupola360 joined ASPEED Technology at COMPUTEX 2026 in a joint booth to showcase how 360° panoramic vision, AI, IoT control, and mobile patrol platforms can come together for smarter remote site management.

This year was especially exciting for us because Cupola360 unveiled something new for the first time: Omni Speed Dome , our next-generation 360° panoramic speed dome.

Unlike traditional mechanical speed dome cameras that rely on physical rotation, Omni Speed Dome is designed to deliver instant 360° panoramic monitoring without panning delays. This helps operators reduce blind spots, keep full situational awareness, and monitor multiple areas on a single screen without losing the bigger picture.

With Omni Speed Dome, teams can view the full scene, zoom in, lock onto key targets, and track important areas more efficiently. It is built for real-world environments where visibility, speed, and context matter.

At COMPUTEX 2026, visitors also explored three demo zones from Cupola360:

Reality Remote Management
A closer look at how panoramic vision, AI, real-time information, and remote control can support distributed site management.
360° Immersive Surveillance
A live demonstration of how full-scene visibility can help teams monitor wide-area environments with more context.
AI Vision for Robotic Patrol
A hands-on view of how Cupola360 cameras and camera modules can be integrated with robots, unmanned vehicles, and mobile patrol platforms.

For Cupola360, COMPUTEX 2026 was not only about showing products. It was a chance to share our vision for Reality Remote Management and connect with partners, customers, and industry professionals exploring the future of AI-enabled remote operations.

Thank you to everyone who visited us at the ASPEED × Cupola360 booth. We look forward to bringing 360° panoramic AI vision into more real-world applications.

Campus Safety Monitoring: How AI Detection and Panoramic Vision Support Safer Schools

Thu, 11 Jun 2026 07:05:42 GMT

Campus Safety Monitoring

Schools are places for learning, growth, and community. From classrooms and hallways to playgrounds, entrances, parking areas, and sports facilities, a campus is often an open environment where students, teachers, staff, visitors, and parents move through the same space every day.

However, campus safety has become an increasingly important concern for many schools. Incidents such as student conflicts, bullying, unauthorized access, teacher-student disputes, vandalism, and emergency situations can happen quickly. In many cases, the challenge is not only how to respond, but how early the school can understand what is happening.

Traditional surveillance systems can help record events, but they often provide fragmented views. A fixed camera may only capture one direction, while critical activity may happen near a hallway corner, entrance, stairway, playground edge, or crowded common area. When an incident occurs, school staff may need to switch between multiple camera feeds to understand the full situation.

This is where wider scene awareness becomes important.

A school campus is not a single controlled space. It includes multiple open and semi-open areas, each with different safety risks. Entrances and exits need to be monitored for unauthorized access. Hallways and stairways may become conflict points during class changes. Playgrounds, sports areas, and parking lots require wider visibility because incidents can happen across a large area.

For campus safety monitoring, visibility is not only about recording video. It is about helping school teams understand the context of an event. Who is nearby? Where did the incident start? Is the situation escalating? Are staff members already responding? These questions are difficult to answer when video coverage is broken into narrow, disconnected views.

360° panoramic vision can help reduce blind spots and give teams a wider view of public campus areas. With broader coverage, staff can better understand what is happening across the scene instead of relying only on isolated camera angles.

Why Campus Safety Requires Full-Scene Visibility

How AI Detection Can Support Faster Response

AI-assisted detection can add another layer of support to campus safety systems. Instead of waiting for someone to manually notice every abnormal situation, AI can help identify selected risk signals and notify staff when attention is needed.

For example, AI object detection may support the identification of dangerous objects such as weapons or sharp objects in public areas. AI can also be configured to detect unusual movement, restricted-area access, crowd gathering, or other abnormal events based on the school’s safety policy and local regulations.

The goal is not to replace teachers, security staff, or administrators. The goal is to help them respond earlier with better information.

When a potential risk is detected, the system can send alerts to authorized staff and provide a live view or shared image of the relevant area. This helps teams quickly understand where the issue is happening, evaluate the situation, and decide the next step.

In urgent moments, faster awareness can make a real difference.

Supporting Incident Review and Accountability

Campus monitoring is not only useful during an event. It can also help after an incident has occurred.

When conflicts, accidents, property damage, or disputes happen on campus, schools often need to understand what actually took place. A wider visual record can help administrators review the timeline, clarify responsibility, and reduce misunderstandings between students, parents, teachers, and staff.

This is especially important in situations where different people remember the event differently. Video context can support a more objective review process and help schools make fairer decisions based on evidence rather than assumptions.

However, this should be handled carefully. Access to footage should be limited to authorized personnel, and incident review should follow clear school policies.

Balancing Safety and Privacy in Schools

Campus safety monitoring must be designed with privacy in mind. Schools are sensitive environments, and monitoring should never feel like constant personal surveillance.

A responsible approach should focus on public and safety-critical areas such as entrances, hallways, parking lots, playgrounds, sports areas, and other shared spaces. Private spaces such as restrooms, changing rooms, counseling rooms, and personal areas should never be monitored.

Schools should also define clear rules for data access, video retention, alert handling, and incident review. The purpose of AI-assisted monitoring should be transparent: to support safety, emergency response, and fair incident review, not to track individual behavior without reason.

Trust matters. A good campus safety system should protect people while respecting the boundaries of privacy.

How Panoramic Vision Supports Safer Campus Environments

Schools need to be safe places for students, teachers, and staff. As campus environments become more complex, safety monitoring should move beyond fragmented camera views and passive recording.

AI detection and 360° panoramic vision can help schools identify risks earlier, understand incidents faster, and respond with better context. At the same time, these systems must be deployed with clear privacy boundaries, responsible access control, and transparent policies.For schools, panoramic vision and AI detection can work together to create a more complete safety monitoring approach. Panoramic cameras can provide wider visual coverage, while AI can help highlight selected risks that require attention.

This combination can support campus teams in several ways:

Reducing blind spots in public areas
Helping staff understand incidents with more context
Supporting faster alerts for potential risks
Improving emergency response coordination
Providing clearer records for post-incident review
Helping schools design safer monitoring coverage

Campus safety is not about watching everyone all the time. It is about making sure that when something dangerous happens, the school has the visibility, context, and response tools needed to act quickly and responsibly.

The future of campus safety is not about more surveillance. It is about better awareness, faster response, and safer learning environments.

Related Resources

Conclusion

Schools need to be safe places for students, teachers, and staff. As campus environments become more complex, safety monitoring should move beyond fragmented camera views and passive recording.

The future of campus safety is not about more surveillance. It is about better awareness, faster response, and safer learning environments.

Swimming Pool Safety Monitoring: Why AI Drowning Detection Needs Full Visibility

Thu, 11 Jun 2026 02:50:59 GMT

Swimming Pool Safety Monitoring: Why AI Drowning Detection Needs Full Visibility

Swimming pools are places for exercise, recreation, and family activities, especially during the summer. Public pools, hotels, schools, sports centers, and community facilities all rely on lifeguards and safety teams to keep swimmers safe.

However, swimming pools can also be high-risk environments. When someone experiences sudden discomfort, cramps, fatigue, or a drowning emergency, the first few seconds are critical. In a crowded pool, a swimmer in distress may not be easy to identify. They may not shout for help, and their movement may look similar to floating, resting, playing, or breath-holding practice.

This is why many facilities are beginning to adopt AI drowning detection and AI-assisted pool monitoring systems. These technologies can help detect abnormal behavior and alert staff when something unusual happens. But AI detection depends on one essential condition: the system must be able to clearly see the areas it is expected to monitor.

AI detection does not begin with the algorithm. It begins with visibility.

If a camera cannot capture the corner of a pool, the area near a ladder, the pool edge, or a crowded zone with overlapping swimmers, the AI has no visual data to analyze. In that case, the issue is not only whether the AI model is accurate. The more fundamental question is whether the scene was fully visible in the first place.

Swimming pools are visually complex environments. Water reflection, ripples, splashing, lane dividers, lighting changes, and partial submersion can all make it harder for both AI systems and human operators to understand what is happening in real time. Even with trained lifeguards on site, it is difficult to monitor every corner of a pool area at every moment.

Why Visibility Matters in AI Pool Monitoring

Common Blind Spots in Swimming Pool Safety Monitoring

Blind spots are one of the biggest challenges in AI-assisted pool safety. A blind spot is not always just a physical area outside the camera frame. In real-world pool environments, blind spots can appear in several ways.

Geometric blind spots occur when corners, pool edges, ladders, children’s areas, or deep-water zones fall outside the camera’s field of view.
Occlusion blind spots happen when a swimmer is blocked by other people, lane dividers, pillars, pool structures, or equipment.
Visual complexity can also affect detection. Water reflection, splashing, ripples, and partial submersion may make a swimmer difficult to recognize clearly.
Behavioral ambiguity is another challenge. In a pool, abnormal behavior may look similar to normal swimming, floating, resting, or playing.
Operational blind spots can happen even after an alert is generated. Staff still need to know where the alert comes from, who should respond, and how quickly action should be taken.

This is why swimming pool safety monitoring should not be evaluated only by AI detection accuracy. Camera placement, field of view, coverage design, and response workflow are just as important.

What Facilities Should Consider Before Deploying AI Drowning Detection

Before deploying an AI drowning detection system, facility teams should first evaluate whether all high-risk areas are clearly visible. This includes pool corners, edges, ladders, entrances, children’s areas, deep-water zones, and areas where swimmers may overlap or gather.

Teams should also consider whether the system provides enough context for staff to respond quickly. A reliable AI-assisted pool monitoring system should support lifeguards with better situational awareness, not create a false sense of security.

AI can help detect potential risks faster, but it should work together with trained personnel, safety procedures, and clear response workflows.

How 360° Panoramic Vision Can Support Pool Safety

This is where 360° panoramic vision becomes relevant.

A wider field of view can help reduce fragmented monitoring and provide a more complete understanding of the poolside environment. Instead of relying only on multiple narrow camera angles, panoramic vision can help teams see more of the full scene with better context.

For AI-assisted pool monitoring, broader visibility can support better coverage planning, incident review, and human-AI collaboration. It does not replace lifeguards, safety procedures, or human judgment. However, it can help teams identify coverage gaps, understand where incidents occur, and improve how monitoring systems are deployed.

Conclusion

AI drowning detection can be an important tool for swimming pool safety monitoring, but it should not be treated as a standalone solution. Its performance depends on what the camera system can see and how the entire monitoring workflow is designed.

For public pools, hotels, schools, sports centers, and recreational facilities, safety monitoring should be built on visibility, context, and human-AI collaboration.

Before AI can detect danger, the system must first see the full scene.

Cupola360 Insights & Stories

不是所有 360° 相機都一樣：如何選擇適合的 Camera Architecture？

選擇相機，應該先理解它是為了解決什麼問題

五種相機架構，其實在解決五種不同的視覺需求

為什麼機內硬體即時拼接對 AI 更重要？

360° 運動相機：完整拍攝，之後再決定怎麼呈現

魚眼相機：從單一位置取得大範圍 Overview

Supporting AI-Enabled Monitoring Across Different Sites

多方向相機：一個安裝點，同時看多個重要方向

PTZ：當任務是把遠方的重要目標看清楚

Cupola360 即時全景相機：完整場景，而且立即可用

比規格更重要的是：影像最後要怎麼被使用？

五種相機，其實是在回答五個不同問題

不只選一種相機，而是建立更完整的視覺系統

結論：相機架構不同，但現場需要的是完整視野

Not All 360° Cameras Are Built the Same: Choosing the Right Camera Architecture

Start With the Problem the Camera Was Designed to Solve

Five Architectures for Five Different Visual Needs

360° Action Cameras: Capture First, Decide the View Later

Fisheye Cameras: Broad Coverage From a Single Viewpoint

Multidirectional Cameras: Several Important Directions From One Device

Supporting AI-Enabled Monitoring Across Different Sites

PTZ Cameras: When the Priority Is Long-Range Detail

More Important Than Specifications: How Will the Video Be Used?

Cupola360: Complete Scene Data, Ready in Real Time

Why In-Camera Hardware Stitching Matters for AI

Five Architectures, Five Different Questions

Beyond One Camera Type: Building a More Complete Visual System

Conclusion: Different Cameras, One More Complete View

從全景視覺到數位孿生：打造 AI 資料中心的遠端維運能力

將 360° 全景影像、設備端感測、熱資訊與電力資料帶進數位孿生，讓資料中心的實體環境變得可視、可判讀，也更容易遠端管理。

先讓資料中心的「實體世界」被看見

把影像、設備狀態與電力資料帶進數位孿生

三方能力如何形成完整的維運視角？

從「看見異常」走向「理解異常」

數位孿生的價值，不只是「做一個 3D 機房」

不取代既有管理系統，而是補上 Physical Visibility

從 Physical Data 建立更主動的資料中心維運

相關資源

關鍵亮點

AI 資料中心的維運挑戰，已不只存在於伺服器內部

AI Data Center Digital Twin | Panoramic Vision and Remote Operations

Connecting panoramic vision and physical data through a digital twin for smarter data center operations.

Making the Physical Data Center Visible

Bringing Video, Equipment Status, and Power Data into the Digital Twin

How the Partner Technologies Work Together

From Seeing an Alert to Understanding It

A Digital Twin Is More Than a 3D Data Center Model

Complementing Existing Management Systems with Physical Visibility

Turning Physical Data into More Proactive Data Center Operations

Related Resources ﻿

Key Highlights

AI Data Center Operations Go Beyond the Server

高風險任務中的全景視野：360° 影像如何支援無人載具遠端勘查

透過 360° 全景影像，讓無人載具、機器狗與移動式巡檢平台在複雜場域中取得更完整的周邊視野。

為什麼無人系統需要 360° 視野？

高風險區域勘查：先看清楚，再決定下一步

基地與關鍵設施巡檢：讓巡檢紀錄保留空間脈絡

Cupola360 如何支援這項應用

相關資源

從移動平台到 Reality Remote Management

遠端態勢感知：讓不同單位進入同一個全景場景

應用亮點

A Wider View for High-Risk Missions: 360° Vision for Unmanned Field Inspection

Using panoramic cameras and AI detection to monitor outdoor areas, reduce blind spots, and respond with better context.

Why Unmanned Systems Need 360° Visibility

High-Risk Area Inspection: See the Scene Before Taking the Next Step

Facility and Critical Infrastructure Patrol: Preserving Spatial Context

How Cupola360 Supports This Application

Related Resources

From Mobile Platforms to Reality Remote Management

Remote Situational Awareness: One Panoramic Scene for Multiple Teams

Application Highlights

從遠端參訪到安全引導：資料中心的 360° 全景影像應用

結合全景相機、機器人、Cupola360⁺ Patrol 與即時 AI，讓參訪導覽與工程師進場流程更可控。

資料中心參訪與工程作業，為什麼需要更可控？

遠端沉浸式參訪：不用到現場，也能理解資料中心

受控工程師引導：機器人帶路，AI 協助確認安全動線

方案架構：機器人、全景影像、平台與即時 AI

從遠端看見，走向遠端理解與管理

Related Resources

Related Resources

Related Resources

Related Resources