本研究已成功整合影像擷取技術、勞工身份辨識設備、資料庫管理技術及網際網路傳輸技術，研發出一套可整合勞工影像監測資料之系統。本系統利用具有紅外線定位能力之E-PAS系統提供勞工身份、區域位置及區域濃度資訊，同時利用影像擷取卡、監控攝影機結合個人電腦構成區域伺服器(Local Server)以進行影像擷取及影像儲存，之後透過網路傳送勞工監測資訊至中央伺服器(Central Server)並可即時估算勞工暴露量。此外，本系統利用分散式架構及事件驅動概念能有效節省影像儲存容量及整合勞工身份、區域停留時間、區域位置、作業影像等暴露資訊。 依據系統基本性能測試結果顯示，本系統目前已可進行4個區域之勞工即時監測並估算勞工暴露量。實驗結果也顯示影像擷取設備可穩定提供每秒影像資訊，其單張影像擷取時間在0.58~ 0.62秒。另外，本系統對於30分鐘的暴露資訊僅需39秒便可透過網路整合影像資料以提供研究人員進行回顧分析。最後本系統於連續5天模擬危害暴露測試實驗中，中央伺服器可正確記錄具有暴露危害的時段，其記錄觀察正確率可高達99％以上，顯示本系統可正確整合影像暴露監測資訊，並提供研究人員進行暴露資訊回顧分析。 An Internet-based Video System(EVS) has been developed by integrating video capture, personal identification, database management, and auto-rendering techniques. The EVS uses multi-camera to capture worker’s activities, and stores images in its local video server when a particular worker entering the predefined working area. Meanwhile, all other exposure related messages, such as worker’s identifications, positions, and exposed concentrations are continuously sent back to the central server through internet. But, in order to reduce the demand of internet bandwidth, video clips are temporally stored in the local server until receiving the retrieving command sending from the central server when a particular exposure event needed to be reviewed. By tracking worker’s position, we were able to retrieve video clips from different local video servers, and integrated them into a movie like video. The gives the research a great tool to improve their efficiency when doing a behavior-based safety or exposure assessment is required. Laboratory results showed that it took only 39 seconds to retrieve 30 minutes videos form local servers for retrospect analysis. The EVS successfully reported 99% of the high exposure events and those workers images in a 5 days no-stop laboratory experiment. These results indicated that the EVS can integrate exposure information correctly, and offer researchers to review and analyze the exposure information.
