Video Image Target Detection And Tracking Based On Mean Shift

With the rapid development of computer science and media technology, the computer vision technology has become one of great concern in various fields. It has important Research value in science and engineering, production and life, which are widely used in intelligent human-computer interaction,medical diagnostics,intelligent robot,video surveillance and other areas.In this paper based on the study of algorithms for detection and tracking of moving objects and on the analyzing of problems in applications,improved methods are proposed to tackle with the problems.In moving target detection, from the algorithm of the concepts, principles, processes and key technologies, we have had a detailed analyze of the most common moving object detection algorithm. Validation and comparison by experiment, then analyzes and summarizes the advantages and disadvantages of each algorithm. Gaussian Mixture Model(GMM) is combined with correlation operation of the background image to automatically update the study factor, then the adaptability and robustness of the background model are significantly increased.In moving target tracking,this article focuses on the problems of complex scene moving object tracking algorithm research and improvement. Firstly, the article makes a in-depth analysis on the Mean Shift tracking algorithm.View of the algorithm exists the shortcoming of when moving object size change track frame fixed lead tracking error and even loss the track object, through adjust the bandwidth of moving target, proposed on a improved algorithm to treat object scale adaptive changes based on Mean Shift, the algorithm in the process of object motion in real-time change the size of the tracking object is adjusted by the track frame such that lock the moving object, to reduce the tracking error.Secondly, in order to solve the occlusion of object motion in complex scenes, the moving object prediction problem has been researched based the Kalman filter. As classic Mean Shift algorithm lack of template update mechanism in the process of tracking object when the object is obscured, design an anti-obscure tracking algorithm by fusion Mean Shift and Kalman filter, in this algorithm, using the changes of the Bhattacharyya coefficient values to determine the time occlusion occurs and occlusion end. The improved algorithm,due to the addition of target template updates mechanism, so to be able to more accurately describe the target feature to improve the accuracy and efficiency of the algorithm.Finally, the proposed algorithm is implemented in Opencv and Matlab,and analysis the experiment results. At the same time,summary this paper work done,and pointed out the existing problems and further research ideas.