Segmentation And Tracing Of Moving Object Based On The Theory Of Statistic Denoise

The recognition and tracing of moving object is a very popular subject in the area of Computer Vision, Image Processing and Pattern Recognition. Although some achievements have been made in the studying of moving object detecting and tracing, how to detect and trace moving object in complex background and meet demand of real-time processing is still hotspot of research in this domain because moving object often stands in the diverse surroundings. This present thesis focuses on the detection and tracing of moving objects in complex background.? Drawn on such technologies and research of moving object detection as optical flow method, frame difference method and background subtraction method, this thesis proposes, based on the theory of statistic denoise, a new method of achieving background. Compared with other methods of achieving image background, this approach can simultaneously achieve and update background. Combined with detection method of bit-layer XOR, it realizes rapid segmentation and detection of moving object. The experiment results show that this method of moving object detection has the advantages of denoise, simple computation and desirable effect and efficiency. ? Method of Kalman predict is discussed in terms of tracing moving object. Parameters of Kalman forecast are designed according to requirements of system and subsequently, system state vector, state transfer matrix and observation matrix are determined. X direction average error of object tracing using Kalman forecast is analyzed, and track and search range are estimated. The experiment verifies the validity of tracing model.