Cooperative hybrid multi-camera tracking for people surveillance

Developing and improving people surveillance in public areas has been a topic of interest over past several years. A novel hybrid visual tracking system for event detection and people tracking is proposed in this paper. Our surveillance system consists of a calibrated stationary and a pan, tilt, and zoom (PTZ) camera. The stationary camera has a wide field of view, and it detects global events. The PTZ camera is triggered once an event is detected, and it pans and tilts to center the target based on the camera geometry. Then, the camera zooms in to obtain levels of details from the target which may not be clear in the wide view of the stationary camera. By applying color-based particle filtering (PF), the active camera is able to continuously centers and tracks the target in high resolution images. Experimental results are carried to further demonstrate the practicality of our proposed event monitoring system.