近幾年來，許多研究報告指出，監視攝影機的建置，有助於預防犯罪的發生及提高破案率，因此，無論是公共場合還是私人空間，我們皆可看到攝影機的身影。但隨著攝影機的普及，伴隨而來的是大量的監控影片，儲存空間及人力的查找成為一個相當重要的問題。
除了協助破案，影片的儲存也是迫切需要解決的問題。按規定，影片需要保存一段時間以供日後調閱，但隨著影片數量增加速度越來越快，空間需求與日俱增。因此，需要一個有效的方法去降低監控影片所需的儲存空間。
在本篇論文中，我們提出了一套能適應不同環境並有效降低儲存空間的監控影片濃縮儲存方法。我們使用高斯混合背景模型、背景相減法取出前景，並用陰影去除方法及型態學運算，提升取得前景品質。接著透過Mean shift演算法及Apperance模型偵測及追蹤移動物體。在擷取移動物體影像時，我們提出了一套移動物體影像前處裡的流程，提升還原影片時的畫質，並提出用影片的格式來儲存移動物體的連續影像。在影片分析方面，我們使用卡爾曼濾波器平滑移動物體軌跡，並提出使用線性回歸修正軌跡資訊，接著使用自我組織網路根據不同的屬性對移動物體進行分析，最後，在影片還原部分，我們根據分析的結果，提供使用者多項還原功能，讓使用者能根據自己的需求，還原不同呈現方式的影片，降低人力查找的時間。
實驗部分，我們針對不同的場景進行分析，我們的方法可以有效追蹤移動物體，並解決移動物體遮蔽問題，影片經濃縮儲存後，檔案大小小於原始檔案的1/6以上。經過我們方法改良的畫質，也被證明有顯著的提升。經修正的軌跡資訊，有助於提升分群結果準確度達12%，整體移動物體分群的準確率可達93%以上。

In recent years, intelligent surveillance system has become an important issue because it helps government and safeguards solve more criminal cases, and hence improves the detection rate. The popularity of surveillance cameras is rising and can be seen in either public or private areas. The huge amount of video brings some issues on searching, storage, and needs a lot of time for detectives to watch.
Beside help solving crimes, the video storage problem is also a pain for the surveillance system industry, According to regulations, such video should be kept for a long period of time, but the number this kind of video is increasing fast. Therefore, an efficient method for magnificently reducing the size of a surveillance video is needed.
In this thesis, a video condensation system is proposed. Unlike other condensation methods, our proposed method focus on reducing storage size. In addition, our proposed method help users de-condense the video to an original video without losing any content. We use Gaussian mixture model to extract foreground objects, and improve the quality of foreground by the use of shadow removal and morphological operations. Then we use mean-shift algorithm and appearance model to track moving foregrounds. In the video analysis part, we use Kalman filter to smooth the trajectory of moving objects, and use linear regression to correct trajectory information. Finally, in the video de-condensation part, we enable user to choose their preferred result to reduce the video length efficiently.
In the experiment part, video under multiple real-environments are evaluated. The result shows that our proposed method is able to track moving objects efficiently, and without the problem of object occlusion. After using our proposed condensation methods, the file size is less than 1/6 times as the original video. The results show that the condensation quality has improved compared to other condensation methods. The corrected trajectory information helps improve the accuracy of analyze result of 12%. The overall moving object trajectory clustering accuracy is more than 93%.

[1]“台北市社區鄰里監視系統在犯罪預防上成效評估之研究”, [Online]Available: https://www.cib.gov.tw/Upload/Files/778.pdf
[2]“台北市政府警察局網站 新建置之監視器成效”, [Online]Available: http://www.tcpd.taipei.gov.tw/ct.asp?xItem=69589798&ctNode=67726&mp=10800G.
[3]“台北市政府警察局新聞稿 郝市長讚許新錄影監視系統犯罪偵防成效優異”, [Online]Available: http://www.tcpd.taipei.gov.tw/ct.asp?xItem=50947067&-ctNode=45192&mp=108001.
[4]P. Nemanja, J. Nebojsa, and H. Thomas, “Adaptive video fast forward,” Multimedia Tools and Applications, pp. 327-344, Netherlands ,2005.
[5]A. Khosla, R. Hamid, L. Chih-Jen and N. Sundaresan, “Large-Scale Video Summarization Using Web-Image Priors,” IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pp. 2698-2705, June. 2013.
[6]A. Rav-Acha, Y. Pritch, S. Peleg, “Making a Long Video Short: Dynamic Video Synopsis,” IEEE Computer Society Conference on Computer Vision and Pattern Recognition, vol. 1, pp.435-441, June. 2006
[7]Y. Pritch, A. Rav-Acha, and S. Peleg, “Nonchronological video synopsis and indexing,” IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 30, no. 11, pp.1971-1984, 2008.
[8]L. Zhuang, P. Ishwar, and J. Konrad. “Video Condensation by Ribbon Carving.” IEEE Transactions on Image Processing,” vol. 18, no. 11, pp. 2572-2583, 2009.
[9]C. Stauffer and W.E.L Grimon, “Adaptive Background Mixture Models for Real-time Tracking,” Computer Vision and Pattern Recognition, vol. 2, pp. 246-252, 1999.
[10]R. Cucchiara, C. Grana, M. Piccardi, and A. Prati, “Detecting Moving Objects, Ghosts, and Shadows in Video Streams,” IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 25, no. 10, pp. 1337-1342, 2003.
[11]R.C. Gonzalez and R.E. Woods, “Digital Image Processing,” Second Edition, Addison-Wesley, Massachusetts, 1992.
[12]R. M. Haralick, “Some neighborhood operations”, Real Time/Parallel Computing Image Analysis, Plenum Press, New York, 1981.
[13]A. Hashizume, R. Suzuki, H. Yokouchi, H. Horiuchi and S. Yamamoto, “An algorithm of automated RBC classification and its evaluation,” Japanese Journal of Medical Electronics and Biological Engineering, vol. 28, no. 1, pp. 25–32, Japan, 1990.
[14]K. Suzuki, I. Horiba, and N. Sugie, “Linear-time connected-component labeling based on sequential local operations,” Computer Vision and Image Understanding, vol. 89 , no. 1, pp. 1-23, 2003.
[15]G. R. Bradski, “Computer Vision Face Tracking for Use in a Perceptual User Interface,” Intel Technology Journal, vol. 2, no.2, pp. 12-21, 1998.
[16]K. Fukunaga, “Introduction to Statistical Pattern Recognition,” Academic Press, Boston, 1990.
[17]D. Comaniciu, V. Ramesh and P. Merr, “Real-time tracking of non-rigid objects using mean shift,” IEEE Conference on Computer Vision and Pattern Recognition, Hilton Head Island, SC, vol. 2, pp. 142-149, Jun. 2000.
[18]D. Comaniciu and P. Meer, “Robust Analysis of Feature Spaces: Color Image Segmentation,” Computer Vision and Pattern Recognition, pp. 750-755, 1997.
[19]A. Senior, A. Hampapur, Y.L. Tian, L. Brown, S. Pankanti and R. Bolle, ”Appearance Models for Occlusion Handling,” Proc. Second IEEE Workshop Performance Evaluation of Tracking and Surveillance, 2001.
[20]M. Basu, “Gaussian-based edge-detection methods-a survey,” IEEE Transactions on Human-Machine Systems, vol. 32, no. 3, pp. 252-260, Aug. 2002.
[21]Gary Bradski and Adrian Kaehler, Learning OpenCV: Computer Vision with the OpenCV Library, O’Reilly Media, pp. 13-140, 2008.
[22]Z. Zhang and R.S. Blum. A categorization of multiscale-decomposition-based image fusion schemes with a performance study for a digital camera application, Proceedings of the IEEE, vol. 87, no. 8, pp. 1315-1326, 1999.
[23]Allen M. Waxman, Alan N. Gove, David A. Fay, Joseph P. Racamato, James E. Carrick, Michael C. Seibert and Eugene D. Savoye, “Color Night Vision: Opponent Processing in the Fusion of Visible and IR Imagery,” Neural Networks, vol. 10, no. 1, pp.1-6, Great Britianin, 1997.
[24]C. Yin and R.S. Blum. “Experimental Tests of Image Fusion for Night Vision,” 8th International Conference on Information Fusion, vol. 1, 2005.
[25]P.J. Burt and E.H. Adelson, “The laplacian parymid as Acompact Image Code,” IEEE Transactions on Communications, vol. 31, no. 4, pp. 532-540, 1983.
[26]F. Rong, T. Qiufen and L.Guanqun, “Non-linear Weighted Multiband Fusion Image Algorithm. Bhooshan,” IEEE Workshop on Electronics, Computer and Applications, pp.449-452, 2014.
[27]B. Sunil and S. Shipra, “A lossy/lossless coding algorithm using histogram,” Advances in Visual Computing, Springer Berlin Heidelberg, pp. 458-464, 2010.
[28]Y. J. Chanu, T. Tuithung and K. Manglem Singh, “A short survey on image steganography and steganalysis techniques,” 3rd National Conference on Emerging Trends and Applications in Computer Science (NCETACS), Shillong, pp. 52-55, 2012.
[29]Ren-Junn Hwang, Timothy K. Shih and Chuan-Ho Kao, “A Lossy Compression Tolerant Data Hiding Method Based on JPEG and VQ,” Journal of Internet Technology, vol. 1, no. 1, pp. 171-178, 2004.
[30]N. Akhtar, S. Khan and G. Siddiqui, “A Novel Lossy Image Compression Method,” Fourth International Conference on Communication Systems and Network Technologies (CSNT), Bhopal, pp.866-870, 2014.
[31]D.S. Taubman and M.W. Marcellin, “JPEG2000: Standard for Interactive Imaging,” Proceedings of the IEEE, vol. 90, no. 8, 2002.
[32]Blue Book, Lossless Data Compression, CCSDS 121.0-B-1, 1997.
[33]T. Basar, “A New Approach to Linear Filtering and prediction problems,” Wiley-IEEE Press, pp.167-179, 2001.
[34]“Linear regression.” [Online]Available: http://en.wikipedia.org/wiki/Linear_reg-ression.
[35]T. Kohonen, “The Self-organizing map,” Proceedings of the IEEE, vol. 78, no. 9, pp.1464-1480, 1990.
[36]“WEKA”, [Online]Available: http://www.cs.waikato.ac.nz/ml/weka/.