在即時系統之中，中央處理器(CPU)的時脈通常不高，如果以高斯混合模型(Gaussian Mixture Model, GMM)做為前景偵測的核心時，CPU的能力將會是一個關鍵。本論文研究以GMM為核心的物體追蹤法在即時系統上的實作，修改前景偵測的演算法來提高效能，並且以SIP做為事件通知的途徑。
在前景偵測時，本論文以部分更新的方式去建立GMM，用以抵抗前景融於背景，且有助於提升效能。為使影像能即時傳送，快速的處理影像跟正確的偵測前景一樣重要，但兩者皆需要強大的CPU支援。在本文中將會對這個問題討論。
當偵測到前景物體之後，系統借由會話發起協議(Session Initiation Protocol, SIP)播出警示(alert)電話。SIP可以穿越NAT，它的這項能力，使得此一智慧監控系統不單是高階產品，對於SOHO族，也能輕鬆的架設於家中而不必擔心NAT的問題。

In real-time embedded system the CPU normally is not too powerful. Therefore, CPU performance will be the bottleneck if using the Gaussian Mixture Model as the motion detection kernel.
We try to implement and research an intelligent surveillance with a real-time system and trigger the alert via SIP protocol and find a different way to use The GMM for motion detection
For motion detection, this thesis will update GMM partially. This algorithm provides better performance. It also provides the ability to assist motion detection.
In order to transport media stream in time, speedy motion detect and correctly motion detect are the same important. There will be detail statement about above issue with the CPU limitation.
After motion detected, system will send a Session Initiation Protocol (SIP) call as an alert. SIP has the ability of NAT traversal. This ability makes this intelligent surveillance not a high end device any more. It becomes easy for SOHO to setup at home without worry about NAT issue.

[1] 陳信國, “The Implementation of Realtime Image Tracking on IP Camera,” 碩士論文, 私立大同大學資訊工程研究所, 2008.
[2] 羅新銘, “Design and Implementation of Image-based Monitoring System,” 碩士論文, 私立明新科技大學精密機電工程研究所, 2010.
[3] 陳家祥, “An Intelligent Surveillance Based on Statistical Guassian Mixture Background Model with Kalman Filter,” 碩士論文, 國立台灣科技大學資訊工程研究所, 2011.
[4] 林韋宏, 謝家華, 郭文德 且 賴岱佑, 數位多媒體系統概論, 文魁資訊股份有限公司, 2009.
[5] C. H. Anderson, P. J. Burt 且 G. S. van der Wal, “Change Detection and Tracking Using Pyramid Transform Techniques,” in Proc.of SPIE, pp. 72-78, 11 December 1985.
[6] J. Rosenberg, H. Schulzrinne, G. Camarillo, A. Johnston, J. Peterson, R. Sparks, M. Handley and E. Schooler, “SIP: Session Initiation Protocol [RFC3261],” Internet Engineering Task Force, 2002.
[7] H. Schulzrinne, S. Casner, R. Frederick and V. Jacobson, “RTP: A Transport Protocol for Real-Time Applications [RFC 3550],” Internet Engineering Task Force, 2003.
[8] Y.-K. Wang, R. Even, T. Kristensen and R. Jesup , “RTP Payload Format for H.264 Video[RFC6184],” Internet Engineering Task Force, 2011.
[9] 郭文賢, “智慧型百萬像素網路攝影機影音串流處理與即時警報訊息傳遞機制之研究,” 碩士論文, 國立中央大學資訊工程研究所, 2008.
[10] wikipedia, “Normal_distribution,” [線上]. Available: http://en.wikipedia.org/wiki/Normal_distribution. [存取日期: 6 2013].
[11] C. Poynton, Digital Video and HDTV, Morgan Kaufmann, 2003, p. 291–292.
[12] wikipedia, “Chroma subsampling,” [線上]. Available: https://en.wikipedia.org/wiki/Chroma_subsampling. [存取日期: 7 2013].
[13] Fu Chang , Chun-jen Chen and Chi-jen Lu , “A linear-time component-labeling algorithm using contour tracing technique,” 2004.
[14] 徐千洋, Linux C函式庫參考手冊, 初版 編輯, 旗標出版股份有限公司, 2001.
[15] 陳正凱, C++ 函式庫精華錄, 8版 編輯, 金禾資訊股份有限公司, 2001.