A fixed single camera is not sufficient for monitoring a wide area. More cameras can be used, but a problem with integrating all of them will arise. In this paper, a monitoring system to detect and track moving objects in an indoor environment using multiple omni-directional cameras is proposed. Objects captured from different cameras can be integrated automatically, and we can add more cameras to enlarge the monitoring range without changing the system architecture. Such a system is currently being applied to a model of a parking lot for detecting the paths of vehicles.
Index Terms:
Omni-directional camera, hyperbolic mirror, surveillance system, object tracking, permutation matrix, homography matrix
Citation:
Jung-Ming Wang, Ching-Ting Tsai, Shen Cherng, Sei-Wang Chen, "Omni-Directional Camera Networks and Data Fusion for Vehicle Tracking in an Indoor Parking Lot," avss, pp.45, 2006 IEEE International Conference on Advanced Video and Signal Based Surveillance (AVSS'06), 2006