簡易檢索 / 詳目顯示

回結果列表
研究生： 李姿慧
Tzu-hui Lee
論文名稱： 移動物件之軌跡比對應於壓縮影像環境
Trajectory Matching of Video Moving Objects in Compressed Video
指導教授： 陳郁堂
Yie-Tarng Chen
口試委員: 方文賢
Wen-Hsien Fang
陳省隆
Hsing-Lung Chen
林銘波
Ming-Bo Lin
學位類別： 碩士
Master
系所名稱： 電資學院 - 電子工程系
Department of Electronic and Computer Engineering
論文出版年： 2009
畢業學年度： 97
語文別： 英文
論文頁數： 56
中文關鍵詞： 監視系統 、軌跡比對 、多物件追蹤 、隱藏式馬可夫模型 、方向碼 、移動向量
外文關鍵詞： Surveillance system, Trajectory matching, Multi-object tracking, Hidden Markov Model, Chain code, Motion vector
相關次數： 點閱：200  下載：2 
分享至:
查詢本校圖書館目錄 查詢臺灣博碩士論文知識加值系統 勘誤回報

    • 近年來智慧型影像監視系統逐漸成為大家關注的研究主題，在影像監視系統中，我們可以使用移動物件的軌跡進行異常事件偵測、人行為分析與影像內容檢索。本論文中，我們實現一個在壓縮的影像監視系統，針對移動物件進行多物件追蹤及軌跡的比對。本論文主要的研究問題包含壓縮影像的物件擷取、多物件追蹤的影像重疊處理及軌跡的分類。首先，我們使用壓縮影像編碼資訊的移動向量追蹤移動物件，並利用匈牙利演算法及物件配對關聯表進行多物件的配對，最後將建立出來的軌跡使用隱藏式馬可夫模型(Hidden Markov Model)進行事件比對與異常事件偵測。另外，我們使用方向碼重新表示軌跡成為隱藏式馬可夫模型訓練的特徵值。實驗結果顯示我們的方法可以有效的針對不同軌跡事件進行分類。

    In recent years, the intelligent video-based surveillance system becomes an important research issue. Moving object trajectory in the video surveillance system can be used to detect abnormal events, analyze human behaviors and support content-based video retrieval. Hence, in this research, we focus on implementing a video surveillance system, which can track multiple objects and perform trajectory matching in the compressed video domain. The major issues in this research include object segmentation from the compressed video, multiple objects tracking with occluded object handling and trajectory classification. First, we adopt motion vector encoded by the compressed video to tack object movements. Then, we use Hungarian algorithm and relation table to match multiple objects. Finally, we apply Hidden Markov Model to match moving trajectory and detect abnormal events. Specifically, direction chain code is used to represent moving trajectory to Hidden Markov Model. Experiment results show that our proposed approach can effectively classify different event trajectories.

    中文摘要I AbstractII Table of contentsIV List of FiguresVI List of TablesVII Chapter 1 Introduction8 1.1 Overview8 1.2 Problem Statements9 1.3 Objective10 1.4 Summary of the Thesis10 1.5 Contributions10 1.6 Organization of the thesis10 Chapter 2 Background11 2.1 MPEG overview11 2.2 Related works15 Chapter 3 Compressed-domain Object Extraction and Trajectory Representation18 3.1 Object Segmentation from compressed video19 3.1.1 Extract Information from Compressed Video19 3.1.2 Noises and Outliers Removal in the Compressed Video21 3.1.3 Cluster and Label Object Mask23 3.2 Multi-object Tracking with Occluded Object Handling25 3.2.1 Compute Object Similarity Metric25 3.2.2 Minimum Weight Matching28 3.2.3Look-up Object Relation Table33 3.3Simplify Motion Trajectory34 3.3.1Vertex Reduction34 3.3.2 Douglas-Peucker Algorithm35 3.3.3 Exponentially Weighted Moving Average35 Chapter 4 Event Classification37 4.1 Trajectory Feature37 4.2 Hidden Markov Model38 4.3 Longest Common Subsequence40 4.4 Dynamic Time Warping42 Chapter 5 Performance Evaluation44 5.1 Object Detect and Tracking Result44 5.2 Trajectory Classify Result47 5.2.1Performance Metric47 5.2.2Event detect experiments result47 Chapter 6 Conclusion52 Appendix A. Program Pseudo Code53 A.1 Multiple Objects Matching Pseudo Code53 A.2 Longest Common Subsequence Pseudo Code54 Reference55

    [1]L. R. Rabiner, “A Tutorial on Hidden Markov Models and Selected Applications in Speech Recognition”, in Proc. IEEE, vol 77, pp. 257-286, Feb. 1989.
    [2]T.H. Cormen, C.E. Leiserson, and R. L. Rivest, “Introduction to Algorithm”, MIT Press, 2nd edition, 1990.
    [3]J. Meng, Y. Juan, and S.F. Chang, “Scene Change Detection in a MPEG Compressed Video Sequence”, in Proc. IS&T/SPIE, Vol. 2419. pp. 14-25, Feb. 1995.
    [4]F. Brémond, and M. Thonnat, “Tracking Multiple Nonrigid Objects in Video Sequences” in IEEE Trans. Circuits and Systems for Video Technology, vol. 8, pp. 575-584, Sep. 1998.
    [5]J. Gilvarry, “Extraction of Motion Vectors from an MPEG Stream”, MD. Dublin City University, 1999.
    [6]E. Keogh, and M. Pazzani, “Scaling up Dynamic Time Warping for Datamining Application”, in Proc. 6th Int. Conf. on Knowledge discovery and datamining, Boston, MA, pp. 285-289, 2000.
    [7]T. Yu., and Y. Zhang, “Motion Feature Extraction for Content-Based Video Sequence Retrieval”, in SPIE Conf. on Internet Imaging Ⅱ, Vol. 4311, pp. 378-388, 2001.
    [8]F. I. Bashir, A. A. Khokhar, and D. Schonfeld, “Segmented Trajectory Based Indexing and Retrieval of Video Data”, in Proc. Intl. Conf. on Image Processing, vol 2, pp. 623-626, Nov. 2003.
    [9]D. Buzan, S. Sclaroff, and G. Kollios, “Extraction and Clustering of Motion Trajectories in Video”, in Proc. Int. Conf. Pattern Recognition, Cambridge, U.K., pp. 521-524, Aug. 2004.
    [10]D. Keysers, T. Deselaers, and H. Ney, “Pixel-to-Pixel Matching for Image Recognition Using Hungarian Graph Matching”, in Proc. DAGM Symp. Pattern Recognition, Vol. 3715, pp.154-162, Oct. 2004.
    [11]D. J. Zhang Jain, “Object Segmentation and Tracking for Surveillance Video”, MD. I-Shou University, Jun. 2005.
    [12]Z. Fu, W. Hu, and T. Tan, “Similarity Based Vehicle Trajectory Clustering and Anomaly Detection”, in Proc. Intl. Conf. on Image Processing, vol 2, pp. 602-605, Nov. 2005.
    [13]H.Y.M. Liao, D.Y. Chen, C.W. Su, and H.R. Tyan, “Real-time Event Detection and Its Application to Surveillance Systems” in Proc. IEEE Int. Symp. Circuits and Systems, pp.509-512, May 2006.
    [14]C. W. Lin, Z.H. Ling, Y. C. Chang, and C. J. Kuo, “Compressed-domain Fall Incident Detection for Intelligent Homecare” in Journal of VLSI Signal Processing System for Signal, Image, and Video Technology, vol 49, pp. 393-408, Jul. 2007.
    [15]G. Bradski., and A. Kaehler, “Learning OpenCV: Computer Vision with the OpenCV Library”, O’Reilly Media, Inc., 1st edition, Oct. 2008.

    QR CODE