現行單一固定式攝影機放置門口、走道或戶外街頭角落等重要位置，監視錄影環境的一舉一動，但由於監視環境空間太大，拍攝目標影像太小，碰到犯罪行為發生時，僅能拍攝到嫌疑犯的身軀，無法提供清晰影像。現行具有PTZ功能之攝影機，雖具有旋轉與放大影像功能，但僅能追蹤某一特定目標物。而吸附式之360度環場攝影機雖可監控追蹤整個環境的目標物，而後驅使PTZ攝影機取得清晰影像，但在某些開放空間場合，不易架設，如：旅館大廳、室外停車場。
基於上述的理由，本研究設計一雙攝影機取像裝置，結合兩種不同功用之攝影機，一為場景攝影機，用來監控整個環境，追蹤所有的目標物，另一為目標物攝影機，具旋轉、放大功能，用來取得目標物之清晰影像。藉由兩支攝影機之擺設位置，與利用特徵點共線之限制條件，設計一演算法，求得雙攝影機影像之對應關係。無需求得目標物在三度空間之位置，也不需事先得知攝影機的參數，直接利用影像處理的技術，自動求得系統兩台攝影機運作所需要的參數，可對多目標物追蹤，同時可取得清晰之目標物特寫鏡頭，並利用所取得之目標物特寫鏡頭，改進追蹤效果。實驗結果證明了本研究所提出之雙攝影機取像裝置是可以用來取得大環境中小目標之清晰影像。

In traditional surveillance systems, a single CCD camera is installed at the entrance of open space. The problem is that moving targets will be too small to be grabbed so that clearer and recognizable images can not be obtained for identification. As to the CCD camera with panning, tiling, and zooming (PTZ) functions, it can only track one single target.
In this thesis, a dual-camera device is designed to solve the aforementioned problems. Two cameras are installed together for different purposes. One camera called ''sense camera'' is used to monitor the whole space and to track multiple targets. The other one called ''object camera'' with PTZ functions is used to grab high quality object images. Using some geometrical properties, the two cameras can be automatically calibrated by point correspondence. Without knowing the 3D locations of objects and the camera parameters, the dual-camera device can not only track multiple targets but also obtain their high quality images in the widespreading open spaces. The tracking process can be easily switched from one target to another when multiple objects appear in the monitoring space. In addition, the tracking performance is also improved by using the detail image information grabbed from the object camera. Experiments were conducted on a wide variety of scenes and the experimental results reveal the validity of our proposed approach.

[1] Z. Zhang, “Flexible Camera Calibration By Viewing a Plane From Unknown Orientations”, International Conference on Computer Vision (ICCV''99) , pp. 666-673, September 1999.
[2] Z. Zhang, “A Flexible New Technique for Camera Calibration”, IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 22, pp. 1330-1334, November 2000.
[3]C. C. Hu, “Construction of Panormic Images for Monitoring Environment”, IPPR Conference on Computer Vision, Graphics and Image Processing (CVGIP), 2004.
[4] R. I. Hartley, “Kruppa’s equation derived form the fundamental matrix”, IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 19, no.2, pp. 133-135,1997.
[5] 李昆霖, ”基本矩陣之最佳化與歐基里德重建”, 交通大學資訊科學研究所碩士論文, 2002.
[6] H. P. Moravec, “Toward Automatic Visual Obstacle Avoidance”, Proc. of 5th International Joint Conference on Artifitial Intelligence, pp. 584, August 1977.
[7] C. Harris and M. Stephens, “A Combined Corner and Edge Detector”, Proc.4th Alvery Vision Conf. Manchester, pp. 189-192, August 1988.
[8] S. M. Smith and J. M. Brady, “SUSAN: A New Approach to Low LevelImage Processing”, International Journal of Computer Vision, pp. 45-78, 1997.
[9] 謝銘倫, “室內場景之特徵點擷取與追蹤”, 交通大學資訊科學研究所碩士論文, 2003.
[10] B. Georgescu, P. Meer, “Point matching under large image deformations and illumination changes”, IEEE Transactions on Pattern Analysis and Machine Intelligence, pp. 674-688, June 2004.
[11] 鄭智明, “雙攝影機校正研究”, 中華大學資訊工程研究所碩士論文, 2004.
[12] J. W. Hsieh, “Fast stitching algorithm for moving object detection and mosaic construction”, Image and Vison Computing 22, pp. 291~306, 2004.
[13] R. Hartley, “In defense of the eight-point algorithm”, IEEE Trans. on Pattern Analysis and Machine Intelligence, vol. 19, pp. 580–593, 1997.
[14] B. Hu, C. Brown, and A. Choi, “Acquiring an environment map through image mosaicking”, Technical Report, Computer Science Department, University of Rochester, 2001.
[15] D. W. Lim, S. H. Choi, and J. S. Jum, “Automated detection of all kinds of violations at a street intersection using real time individual vehicle tracking”, Proceedings. Fifth IEEE Southwest Symposium on Image Analysis and Interpretation, pp. 126-129, April 2002.
[16] 郭大正, “戶外無人管理停車場監控系統”, 中華大學資訊工程碩士論文, 2003.
[17] Y. C. Chung, C. H. Wang, J. M. Wang, S. C. Lin, S. W. Chen, “Integration of Omnidirectional and Movable Cameras for Indoor Surveillance”, IPPR Conference on Computer Vision, Graphics and Image Processing (CVGIP), 2004.
[18] T. Y. Li, C. M. Wang, Y. I. Hung, K. C. Fan, “A Visual Monitoring System Using Wide-Angle Cameras and Pan-Tilt Camera”, IPPR Conference on Computer Vision, Graphics and Image Processing (CVGIP), 2004.
[19] O. Faugeras, Q. T. Luong, and T. Papadopoulo”, The Geometry of Multiple Images : The Laws That Govern the Formation of Multiple Images of a Scene and Some of Their Applications”, March 2001.
[20] M. Pollefeys, “Tutorial on 3D Modeling from Images”, June 26, 2000.
[21] 吳華泰,”整合特徵資訊的區域匹配法研究”，國立台灣大學土木工程研究所碩士論文, 2003.
[21] L. D. Stefano, S. Mattoccia, M. Mola, “A Change-Detection Algorithm Based on Structure and Color”, IEEE Conference on Advanced Video and Signal Based Surveillance, pp.252-259, 2003
[22] F. Chang, C. J. Chen, “A Fast Method for Labeling Connected Components in an image”, IPPR Conference on Computer Vision, Graphics and Image Processing (CVGIP), 2003.
[23] L. M. Fuentes, S. A. Velastin, “People tracking in surveillance applications”, Proceedings 2nd IEEE Int. Workshop on PETS, Kauai, Hawaii, USA,December 9 2001.
[24]劉成祥, “利用視訊資料作人體走勢分析”,中央大學資訊工程研究所碩士論文, 2002.
[25] B. S. Manjunath, P. Salembier, T. Sikora, “Introduction to MPEG-7 Multimedia Content Description Interface”, 2002.
[26] ISO/IEC 15938-3:2001, “Multimedia Content Description Interface – Part 3: Visual”,Version 1.