Video Processing And Moving Object Tracking Of The Intelligent Robot

In recent years, video tracking is a new research direction, which incorporates many technologies like Computer vision, Pattern recognition, Artificial Intelligence, and so on. At present, domestic many tracking technology research are based on the high-performance PC, and Windows system. The embedded platform is used in this paper. Due to the limited hardware resources, many algorithms should be modified to run on the ARM platform.Firstly, in this paper we build the embedded system platform which mainly contains the communication between the development board and PC, the transplantation of the video open source library, and the development of video driver module. Then, on the above platform, frame difference and background subtraction can make robot detect the moving object. After the robot detected the object, the Mean Shift is taken to track the moving object. Based on the detailed discussion of the Mean Shift, the thesis brings in the corner’s feature of the object to improve the tracking algorithm. The algorithm has its flaws. The Kalman filter is introduced to reduce these drawbacks and improve the effectiveness of the tracking. Finally, an algorithm of particle filter is presented. The algorithm uses particle filter to track the object. It uses color of the object as the standard feature, and matching the particle area’s feature with the standard feature. The particle’s area which has the largest weight is regarded as the object. And the experiment results are showed and analyzed.