Moving Target Tracking based on CamShift Approach and Kalman Filter

The surveillance system has been developed for decades. It reduced the crime and protected the lives and properties of people successfully. Some surveillance system are composed of PTZ cameras. Therefore the moving people or target objects could be tracking by the surveillance system. The surveillance system becomes more and more interment. However, there are two problems occurred, target recognition and target shelter, respectively. In this paper, we proposed the method of tracking a moving object for improving the performance of intelligent surveillance system. The method combined CamShift and Kalman filter for tracking the moving object in a complex background or in an occlusion case. The proposed method is very efficient and that is easily implemented in a real-time system. Three major problems of surveillance system design, moving object detecting, moving object tracking and tracking the object in occlusion, have been conquered in the experimental results