在過去數十年裡，大部分的視訊監控系統的運作主要依賴人為的操作，這些系統大都使用閉路電視系統，其主要功能只是消極的錄影存證，並不能主動地提供偵測資訊，然而隨著數位化影音技術的進步，資料儲存設備之價格降低，光學攝影器材的普及和人工智慧技術的日趨成熟，使得智慧型視訊監控系統的發展更為快速且完善，尤其是其自動化與節省人力的優點，更是讓全世界各國重視的發展課題。
　　本論文主要的目的在於提出一個以人類為目標的視訊監控系統，此系統能偵測、追蹤與判斷移動物的行為狀態，為了達到這個目的，我們首先利用背景相減法來擷取出可能移動物的位置，並配合一些影像處理技術與陰影去除的方法來過濾不可能的目標物，一旦目標擷取出來，我們以目標物的大小、顏色和位移等資訊來追蹤物體。最後為了判斷目標物我們引入了高斯混合模型來表示目標物行為的各種狀態，利用此模式我們可以有效地判斷與辨識目標物的各種行為狀態，最後我們以室內環境的人類移動和跌倒行為的影像，來進行系統的實驗，以驗證該系統是可靠且有效的。

In the past decades, most of surveillance systems used infrared rays, radar rays, or microwaves to detect and identify moving targets from monitoring environments, such as airports, station, banks, and restaurants, by manual manner. However, a lot of attentions have been paid to the developing of automatic and intelligent surveillance systems due to the rapid developments of computer and video techniques to monitor your home or your living environments. Intelligent surveillance system can automatically detect, track, and analyze moving objects, including the behaviors of objects and the occurring of unsafe events, and then send warnings message to people without involving any manual efforts.
The main purpose of this thesis is to present a video surveillance system to detect, track, and analyze human behaviors. First of all, we use the technique of background subtraction to detect different moving objects from video sequences. Then, two key features, i.e., object sizes and colors are utilized to track each detected object and its actions. After that, we introduce the theory of Gaussian Mixture Models (GMM) to model the behaviors of objects. According to the parameters of the models, different object behaviors like running, walking, and falling can be successfully recognized and analyzed. Experimental results show that the proposed method offers great improvements in terms of accuracy, robustness, and stability in the analysis of object behaviors.

[1] Wren C. R., Azarbayejani A., Darrell T. and Pentland A. P., “Pfinder: Real-Time Tracking of the Human Body”, IEEE Transactions on Pattern Analysis and Machine Intelligence, Vol. 19, No. 7, pp. 780-785, July 1997.
[2] Haritaoglu I., Harwood D. and Davis L. S., “W4: Real-Time Surveillance of People and Their Activities”, IEEE Transactions on Pattern Analysis and Machine Intelligence, Vol. 22, No. 8, pp. 809-830, Aug 2000.
[3] T. Olson and F. Brill, “Moving Object Detection and Event Recognition Algorithms for Smart Cameras”, In proc. DARPA Image Understanding Workshop, pp. 159-175, May 1997.
[4] R. T. Collins, A. J. Lipton, T. Kanade, H. Fujiyoshi, D. Duggins, Y. Tsin, D. Tolliver, N. Enomoto, O. Hasegawa, P. Burt and L. Wixson, “A System for Video Surveillance and Monitoring”, Tech. Rep., The Robotics Institute, Carnegie Mellon University, 2000. CMU-RI-TR-00-12.
[5] H. Sidenbladh and M. J. Black, “Learning the statistics of people in images and video”, International Journal of Computer Vision. Vol. 54, Issue 1-3, pp. 183-209, Aug.-Oct. 2003.
[6] Haritaoglu, D. Harwood, and L. Davis, “Active Tracker: Surveillance with Active Camera”, http://www.umiacs.umd.edu/users/hismail/ActiveTracker_Outline.htm
[7] C. Anderson, Peter Burt and G. van der Wal, “Change detection and tracking using pyramid transformation techniques”, In Proceedings of SPIE - Intelligent Robots and Computer Vision, Vol. 579, pp. 72-78, 1985.
[8] Beynon M. D., Van Hook D. J., Seibert M., Peacock A.and Dudgeon D., “Detecting Abandoned Packages in a Multi-camera Video Surveillance System”, In Proceedings of IEEE Conference on Advanced Video and Signal Based Surveillance, pp. 221-228 ,July 2003.
[9] Teknomo K., Takeyama Y. and Inamura H., “Frame-Based Tracing of Multiple Objects”, In Proceedings of 2001 IEEE Workshop, pp. 11-18, July 2001.
[10] P. Dempster, N. M. Laird and D. B. Rubin, “Maximum Likelihood from Incomplete Data via the EM Algorithm”, Journal of the Royal Statistical Society. Series B (Methodological), Vol. 39, No. 1, pp. 1-38, 1977.