本研究想要結合彩色影像和深度影像來設計智慧影像監控系統。近年來，智慧型影像監控系統通常都使用彩色影像來自動地偵測可疑事件，像是人員的進出數量、越線和滯留物等偵測功能。當事件發生時，IVS系統會立即的通報安管人員，減少人為疏忽的可能。由於使用彩色影像的關係，在顏色複雜的背景或光影劇烈變化的場景，前景物體不易被分割出來，使得欲偵測物體不完整，造成事件錯誤的判斷。深度資訊可以克服光影變化問題和相似顏色背景的問題，並且輕易地得到完整的前景物件。然而度感測器有偵測距離上的限制，因此本論文的作法是結合彩色資訊和深度資訊來萃取可能的前景物件。除此之外，多層次背景相減法適用於加快本系統的執行速度。在本研究中影像的標籤法還考慮了深度資訊，從實驗結果三維的連通物件法可以解決重疊物件的問題，並且成功地區分在不同距離上的物件。而在事件偵測方面，物件的移動資訊是非常重要的。在本研究中加入深度資訊，使用三維的平面空間來判斷移動的事件偵測。這樣的方式，準確率可以比使用二維彩色影像來得較好。最後進行各種實驗來證明提出的方法是有效的。

This research intends to integrate color images and a depth image to design an Intelligent Video Surveillance (IVS) system. Recently, IVS usually uses color image to autonomously detect suspicious events, such as people entering, leaving and loitering in some considered area. When a suspicious event is detected, IVS can notify security guards immediately to reduce possible neglects of security guards. Color image are easily influenced by light and shadow change. Also, in the foreground extraction process, due to complicated background color, the foreground objects are not easily distinguished from the background. Depth information can overcome the light and shadow change problem and similar color background problem. It can be found that depth information can divide the foreground object easily and completely. However, the depth sensor has a limit in the detection distance. Therefore, our approach is to integrate both color and depth information to extract possible foreground objects. Besides, a multi-layer background subtraction approach is adapted to speed up the execution speed of the system. In this study, the depth information is also included in the labeling approach. From our experiments, such a 3-dimensional connected-component labeling approach can resolve the object overlapping problem and distinguish objects successfully. Finally, in the suspicious event determination, the movements of an object are important information. In this study, by including the depth information, a 3-dimensional plane is considered in movement event detection. It can be found that the accuracy of suspicious event detection is better than that of using only 2-dimensional images. Various experiments are conducted to demonstrate the effectiveness of the proposed approaches.

[1] IVS技術簡介, http://www.rcg.tv/zh_tw/industry-technology/ivs-overview.html.
[2] QUEST, http://njqs.cn/pro.asp?id=304.
[3] S. M. Yoon and H. Kim, “Real-time multiple people detection using skin color, motion and appearance information,” Robot and Human Interactive Communication, pp. 331 - 334, 2004.
[4] S.-Y. Hou, A Comparative Study on Image-Based and Image-Depth-Based Smart Video Surveillance, Master Thesis, National Taiwan University of Science and Technology, 2011.
[5] ASUS, http://www.asus.com/Multimedia/Xtion_PRO_LIVE/.
[6] J. M. Guo, C. H. Hsia, M. H. Shih,Y. F. Liu, J. Y. Wu ,“High speed multi-layer background subtraction,” IEEE International Symposium on Intelligent Signal Processing and Communication Systems, pp. 74 – 79, 2012.
[7] J. M. Guo, Y. F. Liu, C. H. Hsia, M. H. Shih and C. S. Hsu, “Hierarchical method for foreground detection using codebook model,” IEEE Transactions on Circuits And Systems for Video Technology, Vol. 21, No. 6, pp. 804-815, 2011.
[8] Wikipedia, http://en.wikipedia.org/wiki/Mathematical_morphology.
[9] Morphological Image Processing: Basic Concepts, http://ee.lamar.edu/gleb/dip/10-1%20-%20Morphological%20Image%20Processing.pdf.
[10] R. C. Gonzalez and R. E. Woods, Digital Image Processing, Prentice Hall, 3rd Edition, 2007.
[11] C.-M. Li, The Implementation of The Real-Time Video-Audio Teaching Communication System, Master Thesis, National Taiwan University of Science and Technology, 2011.
[12] C.-Y. Yang, Recovery and Character Segmentation of Chinese Warped Document Images, Master Thesis, National Taiwan University of Science and Technology, 2011.
[13] Yen-Chih Ho, A Public Security Surveillance System for Face Detection and Recognition of Event Related Pedestrian, Master Thesis, National Taiwan University of Science and Technology, 2012.
[14] Fei Zhang, Pinyi Ren, Hao Chen, Guobing Li, “Embedded intelligent video surveillance and cooperative tracking system,” International ICST Conference on Communications and Networking, pp. 632-637, 2012.
[15] Jong Sun Kim, Dong Hae Yeom, Young Hoon Joo, “Fast and robust algorithm of tracking multiple moving objects for intelligent video surveillance systems,” IEEE Transactions on Consumer Electronics, pp. 1165-1170, 2011.
[16] Foroughi, H., Aski, B.S.,Pourreza, H.,” Intelligent video surveillance for monitoring fall detection of elderly in home environments,” International Conference on Computer and Information Technology, pp. 219-224, 2008.
[17] Ji Tao, Turjo M, Mun-Fei Wong, Mengdi Wang , Yap-Peng Tan, “Fall Incidents Detection for Intelligent Video Surveillance,” International Conference on Information, Communications and Signal Processing, pp. 1590-1594, 2005.