Research On The Algorithms Of Moving Object Detection And Tracking

Moving target detection and tracking is a hot topic in the field of computer vision, intelligent video surveillance system is the most basic, the most widely used intelligence analysis, is of great theoretical significance and value in use. Paper based on the target video sequence detection and tracking algorithm, and detection and tracking algorithm used has a further improvement and perfection.In moving target detection, background subtraction method used in this paper moving target detection. Surrender faster method is selected on extraction and updating the background. To achieve the process of establishing the background target detection, extraction in the background during target detection using frame difference. After obtaining background, detecting the moving object use background subtraction. Surrender algorithm by using the background update.In the moving target tracking, this in-depth study of the Mean Shift Algorithm and Cam Shift tracking algorithm, and further comparative analysis of the advantages and limitations of the two algorithms. For a succession of Mean Shift algorithm to speed and robustness characteristics, but also for tracking changes in the window of the adaptive algorithm is the focus of this paper. How determine the trend of tracking targets is the key to solving this problem. Achieve the purpose, the paper introduce the active contour model, the Mean Shift algorithm active contour model tracking window as the outline of initialization, initialize the active contour model that achieve the key issues, but also divided the target area is being tracked. Mean Shift Tracking Based on the segmentation of continuous target area, the comparison, the trend of the target area, thus achieving the adaptive changes in the tracking window. Experiments show that the nuclear activities of contour Bandwidth Mean Shift algorithm can achieve adaptive tracking window size changes with the goal of corresponding changes, this change not only the scene for smaller targets, but also for increasing the target beyond the tracking window conditions, and compared to Cam Shift tracking algorithm has better robustness.