在視覺追蹤應用上區域比對法為一種廣被使用的運動偵測方法，且其應用在pan/tilt視覺追蹤架構中之追蹤效果良好。然而一般的區域比對法僅使用單一影像特徵來進行比對，其缺點是當此特徵受到干擾而比對失敗時，將導致整個追蹤任務失敗，因此本論文結合樣版與輪廓兩個影像特徵來描述目標物，利用雙重影像特徵的比對來提高視覺追蹤演算法的強健性。一般固定的輪廓模型只能描述特定形狀之目標物輪廓，而不適用於會變形的物體，因此本論文使用動態輪廓模型來描述目標物輪廓，利用其可隨物體變形而動態改變其輪廓的特性，可描述一非剛體之目標物輪廓。傳統上動態輪廓模型需手動指定初始輪廓，本論文嘗試提出以適應性背景相減法執行運動偵測並搭配邊界描繪法，使動態輪廓模型的初始輪廓能自動產生。此外在伺服控制部分，本論文在視覺回授路徑上加入一g-h filter，來預測移動目標物的位置，改善視覺回授感測器的回授延遲問題，以期提高視覺追蹤之精確度。欲驗證所提方法之可行性，本論文以一實驗室自行建構之pan/tilt視覺追蹤系統進行數項追蹤實驗，結果顯示本論文所提出之方法效果較原有之方法有明顯改善。

　　Among the motion estimation algorithms used in visual tracking applications, the region-based matching method is one of the most popular approaches. However, generally, the region-based matching method only uses one image feature to perform matching. If there is noise in this particular image feature, very likely the region-based matching method may result in a false result so that the visual tracking task will fail. To overcome this difficulty, this thesis integrates the active contour models with the region-based matching method. By performing multiple image feature matching tasks, the visual tracking algorithms will be much more robust.
　　The contour of a rigid body can be represented using a fixed contour model. However, a fixed contour model is not suitable for the target with a deformable shape. To overcome this difficulty, the active contour models are employed in this study. The active contour models can fit the desired features in the image dynamically. Generally, an active contour is initialized approximately around the target by manually placing a set of discrete points that govern the movement of the contour. Therefore, the conventional contour initialization method cannot be used in real-time visual tracking. To overcome this difficulty, an approach that combines the adaptive background subtraction method with border tracing is proposed in this study. Experimental results show that the proposed approach can initialize contour automatically and make it close to the contour of target.
　　For the servo control unit, a g-h filter is added to the vision feedback loop in this study to predict the position of the moving target so that the sensor delay problem of the image sensor can be improved, and hence increase the tracking accuracy. The pan/tilt visual tracking system developed in our Lab is used to verify the effectiveness of the proposed approach. Several visual tracking experiments are conducted. Experimental results indicate that the tracking performance of the proposed approach can indeed improve significantly compared with that of the conventional approach.

[1] J. Hill and W. T. Park, “Real time control of a robot with a mobile camera,” in Proceedings of 9th International Symposium on Industrial Robots, 1979, pp. 233-246.

[2] S. Hutchinson, G. D. Hager and P. I. Corke, “A tutorial on visual servo control,” IEEE Trans. on Robotics and Automation, vol. 12, pp. 651-670, 1996.

[3] P. Anandan, “Measuring visual motion from image sequences,” COINS Dept. Univ. of Massachusetts, Tech. Rep. COINS- TR-87-21, 1987.

[4] D. Kragic and H. I. Christensen, “Cue integration for visual servoing,” IEEE Trans. on Robotics and Automation, vol. 17, pp. 18-27, 2001.

[5] J. Triesch and C. von der Malsburg, “Democratic integration: Self-organized integration of adaptive cues,” Neural Computation, vol. 13, pp. 2049-2074, 2001.

[6] S. Birchfield, “An elliptical head tracker,” in 31st Asilomar Conference on Signals, System, and Computers, 1997, pp. 1710-1714.

[7] M. Kass, A. Witkin and D. Terzopoulos, “Snake: Active contour models,” International Journal of Computer Vision, vol. 1, pp. 321-331, 1987.

[8] 陳淑嬌，梯度向量流主動輪廓模型於子宮頸抹片細胞邊界偵測之應用，中興大學碩士論文，2003。

[9] 何易展，細胞顯微影像之分割、追蹤與運動分析，成功大學碩士論文，2002。

[10] F. Leymarie and M. D. Levine, “Tracking deformable objects in the plane using an active contour model,” IEEE Trans. on Pattern Analysis and Machine Intelligence, vol. 15, pp. 617-634, 1993.

[11] F. Mendels, C. Heneghan and J. P. Thiran, “Identification of the optic disk boundary in retinal images using active contours,” in Irish Machine Vision and Image Processing Conference, 1999.

[12] J. R. Kim and Y. S. Moon, “Automatic localization and tracking of moving objects using adaptive snake algorithm,” in Information, Communications and Signal Processing, 2003 and the Fourth Pacific Rim Conference on Multimedia. Proceedings of the 2003 Joint Conference of the Fourth International Conference, 2003, pp. 729-733.

[13] W. Kim and Ju-Jang Lee, “Vision tracking using snake for object’s discrete motion,” in Proceedings of the 2001 IEEE International Conference on Robotics and Automation, 2001, pp. 2608-2613.

[14] 陳佩穎，以新型動態輪廓技術完成可追蹤任意形狀物體之強健影像伺服系統，臺灣大學碩士論文，2003。

[15] M. Nemani, T. C. Tsao and S. Hutchinson, “Multi-rate analysis and design of visual feedback digital servo-control system,” Trans. of the ASME, Journal of Dynamic Systems, Measurement, and Control, vol.116, pp. 47-55, 1994.

[16] P. M. Sharkey and D. W. Murray, “Delay versus performance of visually guided systems,” IEE Proceedings- Control Theory and Applications, vol. 143, pp. 436-447, 1996.

[17] T. P. Sim, G. S. Hong and K. B. Lim, “Multirate predictor control scheme for visual servo control,” IEE Proceedings- Control Theory and Applications, vol. 149, pp. 117-124, 2002.

[18] L. D. Cohen, “Note on active contour models and balloons,” CVGIP Image Understanding, vol. 53, pp. 211-218, 1991.

[19] C. Xu and J. L. Prince, “Snakes, shapes, and gradient vector flow,” IEEE Trans. on Image Processing, vol. 7, pp. 359-369, 1998.

[20] R. G. N. Meegama and J. C. Rajapakse, “NURBS snakes,” Image and Vision Computing, Vol. 21, pp. 551-562, 2003.

[21] B. K. P. Horn and B. G. Schunck, “Determine optical flow,” Artificial Intelligence, vol. 17, pp. 285-304, 1981.

[22] D. Murray and A. Basu, “Motion tracking with an active camera,” IEEE Trans. on Pattern Analysis and Machine Intelligence, vol. 16, pp. 449-459, 1994.

[23] C. Wren, A. Azarbayejani, T. Darrell and A. Pentland, “Pfinder: Real-time tracking of the human body,” IEEE Trans. on Pattern Analysis and Machine Intelligence, vol. 19, pp. 780-785, 1997.

[24] R. T. Collins, A. J. Lipton, T. Kanade, H. Fujiyoshi, D. Duggins, Y. Tsin, D. Tolliver, N. Enomoto, O. Hasegawa, P. Burt and L. Wixson, “A system for surveillance and monitoring,” The Robotics Institute, Carnegie Mellon University, Pittsburgh PA, Tech. Rep. CMU-RI-TR-00-12, 2000.

[25] R. C. Gonzalez and R. E. Woods, Digital Image Processing. New Jersey: Prentice-Hall, 2002.

[26] 楊武智，影像處理與辨認。台北：全華，1994。

[27] 王俊凱，基於改良式適應性背景相減法與多重影像特徵比對法之多功能即時視覺追蹤系統之設計與實現，成功大學碩士論文，2004。

[28] M. Sonka, V. Hlavac and R. Boyle, Image Processing, Analysis, and Machine Vision. Pacific Grove: PWS, 1999.

[29] A. A. Amini, S. T.ehrani and T. E. Weymouth, “Using dynamic programming for minimizing the energy of active contours in the presence of hard constraints,” in Proceedings, Second International Conference on Computer Vision, 1988, pp. 95-99.

[30] D. J. Williams and M. Shah, “A fast algorithm for active contours and curvature estimation,” CVGIP Image Understanding, vol. 55, pp. 14-26, 1992.

[31] K. M. Lam and H. Yan, “Fast greedy algorithm for active contours,” Electronics Letters, vol. 30, pp. 21-23, 1994.

[32] E. Brookner, Tracking and Kalman Filtering Made Easy. New York: John Wiley & Sons, 1998.

[33] A. J. Lipton, H. Fujiyoshi and R. S. Patil, “Moving target classification and tracking from real-time video,” in Proceedings of the Fourth IEEE Workshop on Application of Computer Vision, 1998, pp. 8-14.

[34] 陳俊壬，即時視覺伺服追蹤系統之運動偵測與估算，成功大學碩士論文，2003。

[35] 郭勝凱，在複雜背景下以強健視覺伺服為基礎可自動調變倍率之頭部追蹤系統， 臺灣大學碩士論文，2002。

[36] 陳寬益，即時視覺伺服追蹤系統之設計與實現，成功大學碩士論文，2002。