隨著MPEG4視訊標準發展，視訊物件平面觀念也愈來愈受重視。由於每一張畫面都可視為多個物件平面組合，在編碼之前，必須將視訊切割成一連串視訊物件平面。物件運動會造成局部像素強度改變，此改變為影像切割重要資訊，然而攝影機變焦取像時動作則會造成畫面中像素全面性之運動，稱為全域運動(Global motion)，近年來多位學者為排除攝影機造成全域運動，已發展了許多使用運動參數消除攝影機動作造成全域運動之運動補償法。本文提出一套有效率運動補償之仿射運動模型，能快速地解出運動模型中參數，進行運動補償。再結合了時間與空間資訊影像分割方法，進行物件擷取。
本文是以個人電腦為平台建構即時影像追蹤系統為系統架構，以單一固定之廣角鏡頭攝影機監控整個環境。再利用影像相減法，偵測出所有非背景之物體，將所得訊息傳給長鏡頭攝影機組，使監控系統得以同時追蹤多個目標物。首先將長鏡頭所拍攝到之影像做出畫面之間運動比對。由於長鏡頭在拍攝時會有鏡頭拉近或推遠(zoom)、水平移動(pan)、垂直移動(tilt)，將造成之所拍攝影像之全域運動，在擷取物體之前，以一仿射運動模型估計出攝影機移動所造成畫面間之全域運動，經由運動補償，尋找出影像變化資訊，並找出移動物體之範圍。再利用SRG(Seeded Region Growing)演算法，進行細部切割，以背景種子(Background Seed)進行侵蝕，排除範圍內多餘之背景，以求得更精確物體邊緣。在擷取物件後，以廣角鏡頭之背景，計算出長鏡頭所拍攝位置背景影像，並擷取出部份影像，結合物件切割之結果，還原出長鏡頭所拍攝影像，減少影像冗餘之部份，降低影像位元率(Bit rate)，以達到資料壓縮之目的。

With the development of the MPEG4, the concept of video object plane (VOP) is more and more important. Before the MPEG4 video coding, it requires a prior process to decompose the video sequences into several video segmented object planes. As moving objects cause local intensities changes, these changes provide the most important information of the segmentation. Whereas the camera motion also changes intensities, we have to remove the changes caused by camera moving.

The experimentation is using a fixed and single global lens CCD to detect the objects which are not belong to the environment. When global lens CCD finds the objects, the system will pass the objects information to PTZ lens CCD, which is able to capture the better image of the objects. When PTZ lens CCD is tracing objects, CCD`s moving causes global motion between frames. As extracting the objects in these frames, we use an efficient affine model to remove the image intensities changes caused by PTZ lens camera moving. After motion compensation, find the changes due to moving objects, and select the region of the objects. On purpose to find the precise contours of the objects, we use SRG algorithm to erode the background inside the objects region, and finally extract the objects. In the last step, we return back the image of the PTZ lens CCD with global lens CCD`s background and the extracted objects. By this method, we are able to reduce the redundancy and lower the data translation bit rate.

摘要 1
目錄 3
第一章 相關研究 5
1.1 MPEG-4關於物件切割技術之討論 5
1.2 物件切割 8
1.2.1 空間域分割 9
1.2.2 時間域分割 13
1.2.3 結合空間域與時間域分割結果 15
1.3 全域運動與仿射運動模型 20
第二章 研究方向與步驟 25
2.1 研究動機 25
2.2系統架構 26
2.3 物體追蹤 28
2.3.1 廣角鏡頭攝影機影像之擷取 29
2.3.2 長鏡頭攝影機物體之追蹤 34
第三章 影像物件擷取 38
3.1 仿射運動模型與運動補償 40
3.2 計算出物體區域 45
3.3 SRG(Seeded Region Growing)演算法侵蝕背景影像 46
3.4 擷取廣角鏡頭之背景影像 48
第四章 結論與未來工作 54
參考文獻 56

[1] Touradj Ebrahimi , Fernando Pereira, “ The MPEG-4 Book, ” 1st edition ,July 10, 2002.
[2] Abdul H. Sadka , John Wiley & Sons , LTD“ Compressed Video Communications,” Feb.2001.
[3] T. Sikora, S. Member, “The MPEG-4 Video Standard Verification Model,” IEEE Trans. Circuits. Systems. Video Technology, vol. 7, no. 1, pp. 281-298, 2000.
[4] C. Gu, M. C. Lee, “Semiautomatic Segmentation and Tracking of Semantic Video Objects,” IEEE Trans. Circuits. Systems. Video Technology, VOL. 8, NO 5, pp. 572-584, SEPTEMBER 1998.
[5] Dong-Keun Lim, Yo-Sung Ho, “Image Segmentation using Hierarchical Meshes,” Proceedings of the IEEE International Conference on Image Processing, ICIP-99, Kobe, Japan, October, 1999.
[6] L. Vincent ans P. Soille, “Watershed in digital space: An efficient algorithm based on immersion simulations,” IEEE Trans. Pattern Analysis Machine Intell., Vol. 13, pp 583-598, June, 1991.
[7] J. Canny, “A Computation Approach to Edge Detection,” IEEE Trans. On Pattern Analysis Machine Intell. Vol. 8 No. 6, pp. 679-698, 1986.
[8] Deuk Kim, Jaeyoun Yi, Hyun Mun Kim, and Jong Beom Ra, Member, “A Deblocking Filter with Two Separate Modes in Block-Based Video Coding Sung,” Circuits and Systems for Video Technology, IEEE Transactions on , Volume: 9 Issue: 1 , pp. 156 -160, Feb. 1999
[9] D. Wang, “Unsupervised Video Segmentation Based on Watersheds and Temproal Tracking,” IEEE Trans. Circuits. Systems. Video Technology, VOL. 8, NO. 5, pp.539-546, SEPTEMBER 1998.
[10] M. Kim, J. G. Choi, D. Kim, H. Lee, M. H. Lee, C. Ahn, and Y. S. Ho ,“A VOP Generation Tool : Automatic Segmentation of Moving Objects in Image Sequences Based on Spatio-Temporal Information,” IEEE Trans. Circuits Systems. Video Technology, VOL. 9, NO 8. pp. 1216-1226, DECEMBER 1999.
[11] N. Paragios, G. Tziritas, “Adaptive detection and localization of moving objects in image sequences,” Department of Computer Science, University of Crete , Greece , October 1996
[12] Valette, S.; Magnin, I.; Prost, R. “Active mesh for video segmentation and objects tracking,” Image Processing, 2001. Proceeding. 2001 International Conference on , Volume: 2 , 7-10 Oct 2001 pp. 77 -80 vol.2,1998
[13] Gokcetekin, M.H.; Harmanci, M.D.; Celasun, I.; Tekalp, A.M. “Mesh based segmentation and update for object based video,” Image Processing, 2000. Proceedings. 2000 International Conference on , Volume: 1 pp. 343 -346 vol.1,2000
[14] Thomas Meier and King N Ngan, “Automatic Segmentation of Moving Objects for Video Object Plane Generation," IEEE Transactions on Circuits and System for Video Technology, VOL. 8, NO. 5, Sep 1998.
[15] F. T. Azar, and K. E. Tait, “Image Recovery Using the Anisotropic Diffusion Equation,” IEEE Trans. Image Processing, VOL. 5, No. 11, pp. 1573-1579, NOVEMBER 1996.
[16] P. Salembier, M. Pardas “Hierarchical Morphological Segmentation for Image Sequence Coding,” IEEE Transactions on Image Processing, Vol. 3, No. 5, pp.639-851, September 1994.
[17] L. Shafarenko, et. al., “Automatic Watershed Segmentation of Randomly Textured Clolor Images,” IEEE Transactions on Image Processing, Vol. 6, No. 11, pp.1530-1543, July, 1997.
[18] A. Shiji, N. Hamada, “Color Image Segmentation method using Watershed Algorithm and Contour Information,” Proceedings of the IEEE International Conference on Image Processing, ICIP-99, Kobe,Japan, October, 1999.
[19] J. Heuer And A. Kaup,“Global motion estimation in image sequences using robust motion vector field segmentation,” Proceedings of the seventh ACM international conference on Multimedia , pp. 261-264, October 1999
[20] T.Aach, R. Meater, “Statistical model-based change detection in moving video,” Signal Processing, 31, pp. 165-180, 1993.
[21] Smolic, A.; Ohm, J. R., ”Robust global motion estimation using a simplified M-estimator approach ,”Image Processing, 2000. Proceedings. 2000 International Conference, Volume: 1 , 2000.
[22] M. Biering, “Displacement by hierarchical block matching,” in Proc. SPIE Visual Communications and Image Processing(VCIP’ 88), Cambridge, MA, vol. 1001, pp. 942-951 , Nov. 1988.
[23] A. Tremeau, P. Colantoni, “Regions Adjacency Graph Applied to Color Image Segmentation,” IEEE Trans. Image Process., vol. 9, pp. 735-744, April 2000.
[24] Z. Wu, R. Leahy, “An Optimal Graph Theoretic Approach to Data Clustering: Theory and Its Application to Image Segmentation,” IEEE Trans. Pattern Anal. Machine Intell., vol. 15, pp. 1101-1113, Nov. 1993.
[25] N. M. Namazi, P. Penafiel, C. M. Fan, “Nonuniform Image Motion Estimation Using Kalman Filtering,” IEEE Trans. Image Process. Vol. 3, Sept. 1994.
[26] Rolf Adams, Leanne Bischof, "Seeded region growing," IEEE Trans. on PAMI, Vol. 16, No. 6, pp. 641 -647, June 1994.
[27] Andrew Mehnert, Paul Jackway, "An improved seeded region growing algorithm," Pattern Recognition Letters, Vol. 18, , pp. 1065-1071 ,1997.
[28] A. Tremeau, P. Colantoni, “Regions Adjacency Graph Applied to Color Image Segmentation,” IEEE Trans. Image Process., vol. 9, pp. 735-744, April 2000.
[29] A. Neri, S. Colonnese, G. Russo, P. Talone, “Automatic Moving Object and Background Separation,” Signal Processing, vol. 66, issue 2, pp. 219-232, April 1998.
[30] L. Shafarenko, et. al., “Automatic Watershed Segmentation of Randomly Textured Clolor Images,” IEEE Transactions on Image Processing, Vol. 6, No. 11, pp.1530-1543, July, 1997.
[31]　A. Shiji, N. Hamada, “Color Image Segmentation method using Watershed
Algorithm and Contour Information,” Proceedings of the IEEE International Conference on Image Processing, ICIP-99, Kobe, Japan, October, 1999.