Research On Three-Frame Difference Detection Algorithm And Meanshift Tracking Algorithm For Video Moving Target

Computer vision is a challenging subject in the field of Engineering Science, and the detection and tracking of moving objects is a key technology in the research. And this technology has been widely used in the fields of intelligent transportation, medical diagnosis, military engineering and so on. Therefore, it is of great theoretical significance and practical value to further explore and practice the subject.In this paper, the shortcomings of the current moving target detection and tracking algorithms are presented. The existing target detection algorithms and target tracking algorithms are optimized and improved,and the effectiveness of the algorithm is verified by the experimental results.In terms of target detection algorithm, the existing detection algorithms are summarized, and the advantages and disadvantages of each algorithm are analyzed. In this paper, the three frame difference algorithm based on inter frame difference algorithm is deeply studied,and the optimization is made on the basis of the original three frame algorithm. In order to detect the moving target more accurately, the algorithm is based on the three frame difference algorithm and the optical flow method. After a series of experiments, it is proved that the proposed algorithm can accurately detect the target and the feasibility and robustness of the algorithm.About the aspect of moving target tracking algorithm, it mainly discusses the characteristics and defects of MeanShift target tracking algorithm. According to this algorithm, two important disadvantages are put forward, and two improvements are made. On the one hand, through the experimental analysis of the MeanShift algorithm, algorithm of tracking window in the process of implementation of the size does not change according to the target size, which will lead to the error of the tracking results is relatively large or even lost target results. In order to solve this problem, this paper proposes an algorithm that can adjust the tracking window scale in real time according to the amount of tracking information. On the other hand, if the moving speed of the target is faster or the target is occluded, the tracking window will deviate from the target and even lose the tracking target. One of the solutions to this problem is that when the target is occluded, the Kalman filter algorithm is used to predict the target position, so as to realize the real-time update of the target model. Judging whether the target is blocked is based on the change trend of Bhattacharyya coefficient. The experimental results show that the algorithm can deal with the occlusion problem and improve the tracking accuracy of the traditional mean shift algorithm.