Real-time multi-target visual tracking using active camera

Visual surveillance in a dynamic environment has drawn a great deal of attention nowadays. It spans a wide research spectrum, such as for access control, human or vehicle detection and identification, detection of anomalous behaviors, crowd statistics or congestion analysis, human-machine interaction in an intelligent space, etc. Since the field of view of one camera is limited, a camera platform is usually equipped with degrees of freedom to extend its observing range. In order to construct a wide-area surveillance system economically, a camera must be utilized to track multiple targets within its limited field of view. With the growing of computational power and advances of tracking theory, now we can monitor gradually more targets with a single camera simultaneously. Many problems in science require estimation of the state of a system that changes over time using a sequence of noisy measurements made on the system. Bayesian filter provides a rigorous general framework for dynamic state estimation problems. Visual tracking is also one kind of this problem. The sensed sequence of 2D image data, which may lose some 3D information and is usually noisy, includes the target, similar objects, and cluttered background, etc. Here, the visual tracker is proposed to overcome the challenging tracking problems and result in successful multi-target tracking when the objects interact in a group. On the other hand, there is another issue called multi-camera cooperation which attracts increasing attention lately. The overall surveillance capability of the entire set of cameras will not be effectively utilized with poor cooperations, and tracking of some targets will be rather vulnerable. Through the designed strategies of camera task assignment and camera action selection, the distributed camera agents will tightly collaborate with one another. Integrating the multi-target tracking technique and multi-camera cooperative methodology, the prospect of seamless tracking can be realiz- ed stage by stage. One can show that, after the overall system is equipped with the former capability, then we can develop a powerful wide-area surveillance system with near real-time performance.