開放空間上移動物體追蹤之雙PTZ攝影機系統的研究

The Study of a Dual-PTZ-camera System for Visual Tracking of a Moving Target in an Open Area

陳韋毅(Wei-Yi Chen)

指導教授 ： 廖珗洲

朝陽科技大學/資訊學院/資訊工程系/碩士(2009年)

摘要

監視系統使用主動式攝影機，亦即PTZ攝影機，除了能夠監視一個大範圍的區域，更能夠透過光學變焦鏡頭，得到目標物體的細部資訊，如臉部特徵、衣著等等，因此主動式攝影機的應用越來越受到重視，許多研究都將其納入各自的系統當中，形成一種新的多攝影機系統。在本篇研究當中，將提出一個新的雙PTZ攝影機監視系統，稱為Eagle-Eye系統，其概念為仿造生物界老鷹的眼睛構造而得，透過這樣的系統，能夠監視一個開放性的戶外廣場，並且同時得到目標物體的細部資訊，而為了能夠持續的追蹤廣場上的物體，在本篇研究之方法當中使用了兩個追蹤的方法，一種是影像處理的物體偵測，第二種是特徵比對，兩種方法在不同情況下會相互輔助直到追蹤任務結束。在實驗中我們分別對兩個廣場上的移動物體進行隨機取樣追蹤，結果顯示有90%的追蹤成功率；此外，針對移動物體交錯之情形，系統正確追蹤目標物體的成功率為83%，並且系統在夜晚也具備80%的物體追蹤能力。這些結果說明了本篇研究所提出之Eagle-Eye系統適合用於開放性廣場上的物體追蹤。

關鍵字

特徵比對 ； 監視系統 ； 主動式攝影機

並列摘要

An active camera, i.e., pan-tilt-zoom camera, can be use to monitor a wide area and the high resolution of extracted object. Therefore, it is popularly used in recent studies related to intelligent surveillance systems. A novel dual–PTZ-camera surveillance system called “Eagle-Eye” is proposed in this research. The view of the system is designed to imitate an eagle’s eye. That is, the system can keep monitoring the whole area while focusing on the detail of a target. Two tracking techniques, image processing and feature matching, are used alternatively to finish a tracking task. In the experimental results, the ratio of tracking tasks finished successfully for various moving speeds in two different sites is 90 percentages. Besides, the ratio of tasks finished successfully containing occlusion situation is 83 percentages. Furthermore, the successful ratio of tracking tasks in nighttime is 80 percentages. These results show that Eagle-Eye is feasible for tracking objects in an open area.

並列關鍵字

PTZ camera ； feature match ； surveillance system

參考文獻

[1]L. Marchesotti, L. Marcenaro, and C. Regazzoni, “Dual Camera System for Face Detection in Unconstrained Environments,” Proceedings of 2003 International Conference on Image Processing(ICIP 2003), Vol. 1, 14-17 Sep., 2003, pp. 681-684.

[2]R. Bodor, R. Morlok, and N. Papanikolopoulos, "Dual-Camera System for Multi-Level Activity Recognition," Proceedings of 2004 IEEE/RSJ Internation Conference on Intelligent Robots and Systems (IROS 2004), 28 Sep.-2 Oct., 2004, Vol. 1, pp. 643- 648.

[3]C. Micheloni, E. Salvador, F. Bigaran, and G. L. Foresti, "An Integrated Surveillance System for Outdoor Security," Proceedings of IEEE Conference on Advanced Video and Signal Based Surveillance, 15-16 Sep., 2005, pp. 480-485.

[5]S. Khan, O. Javed, Z. Rasheed, and M. Shah, “Human Tracking in Multiple Cameras,” Proceedings of Eighth IEEE International Conference on Computer Vision (ICCV 2001), 2001, Vol. 1, pp.331-336.

[7]C. Micheloni, G. L. Foresti and L. Snidaro, “A network of co-operative cameras for visual surveillance,” IEE Proceedings of Vision, Image and Signal Processing, 8 April, 2005, Vol. 152, pp. 205-212.