國立交通大學 NCTU

多攝影機監控網路拓撲之自動學習技術

2017/01/06
英文名稱Automatic Learning of Spatio-Temporal Relationships in a Camera Network
技術編號Ⅰ08-YPH-11-2
參與教授洪一平教授
所屬計畫VBIE 以視覺為基礎之智慧型環境的建構四年計畫第二期(2008/11~2012/10)
應用產業監控與保全業
技術分類中控相關功能
技術功效提升系統整合性擴大監控範圍
技術狀態可移轉技術

In multi-camera surveillance systems, it is difficult to integrate the results of video analysis of different cameras, especially when the view field of cameras are non-overlapping. This technique is used to establish the spatio-temporal relationships among cameras (including the connection between entry/exit zones of different cameras and its transition time probability distribution) by analyzing the result of single camera tracking. The established relationships allow long-term tracking in a camera network.