本研究已成功的整合影像擷取技術、勞工身份辨識技術及網際網路傳輸技術，開發出一套可應用於勞工作業環境之多點影像監視系統。本系統利用具有紅外線定位能力之TEWAS系統提供勞工身份、位置資訊；利用USB介面之數位攝影機結合個人電腦進行影像擷取；透過網路傳送定點即時影像與勞工定位資訊至後端伺服器；自行開發之伺服器資訊整合程式可根據勞工身份判讀結果，自動整合作業人員身份、停留時間、停留地點、作業影像等資訊，並儲存於資料庫中。研究人員可利用影像解析程式在終端機前取得所有勞工之暴露資訊。 根據系統基本性能測試結果顯示，本系統已可進行三個操作定點即時監測。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0.76 ~ 0.87秒。且影像接收正確率為100 %。另外於實驗室操作觀察實驗中發現，利用本系統影像回顧解析方法所得各操項目之操作時間記錄與傳統人工觀察方法之比較，兩者之相關性屬高度相關(R2值在0.98 ~ 0.99，p > 0.05)。顯示利用本系統界定勞工操作動作與解析操作時間，與傳統人工觀察法之結果並無顯著差異。 An auto-tracking-video-monitoring-system(ATVMS) has been developed by integrating image capture, personal identification, and internet techniques. The ATVMS detects worker’s identity and his/her current location by using an infrared-based positioning system, called TEWAS. Exposure and position data stream are sent back to a central server located in a control room. Meanwhile, a PC-based camera, located at the spot as the worker does, is sending pictures of the worker’s movement back to the same server. By using this technique, motion pictures taken from different sites can be linked together as one movie for reviewer to further investigate the causes of exposures of that particular work. The ATVMS, at present, has the ability to integrate information taken from three different locations. The laboratory tests showed that PC Capture can steadily provide 1.25 frames per second. Images taken by cameras can reach the data server without any frame drop. In two laboratory experiments, worker’s exposure durations in doing particular task were analyzed by direct observation and by reviewing video clips taken by the ATVMS. The results showed that there were no significant differences between these two methods. (R2 = 0.98 ~ 0.99，p > 0.05). The results indicated that the ATVMS can be a promising tool in exploring worker’s exposure patterns under multi-locations.
