The Research Of Moving Target Detection And Tracking Method Based On Stereo Vision And Improved Particle Filtering Algorithm

With the development and progress of the times, computer science has become a direction of mass human unremitting efforts. Moving target detection and tracking is a part of the field of computer science, which is also one of the focus of the human visual research subject and covers the image processing, artificial intelligence, computer and other advanced science and technology.Considering that it is the most critical link in video processing,so if the moving target that is extracted is inaccurate or incomplete, it will produce very big effect to the final results,and even lead to a conclusion contrary to the fact.Although the existing detection and tracking algorithms are emerge in endlessly, there are obvious shortcomings.For example, the target canbe extracted only when it moves in the normal condition,and it is hard to detection and tracking moving targets under the background of chaos.Besides,it is also hard to track target accurately when the target is covered,etc. For a long time, a lot of solving algorithms are constantly put forward to solve these problems and methods by a national scholars.However,the results are not significant.To solve the above problems,this article respectively uses binocular stereo vision technology to detect moving targets,and the improved particle filter technology combining ABCshift algorithm to track moving targets in image.In the aspect of target detection,the research differences in stereo vision image by using the regional correlation matching technology.It effectively avoids the problem of shade effect on target detection,and detects the moving object in the right in the end.In the aspect of tracking,this article makes some adjustment and optimization in the target model selection and similarity measure of particle filter. After the contrast,the improved particle filter method is better in terms of target tracking,the operation steps is less than before and the operation speed is faster. By combining the improved particle filter with ABCshift algorithm,this article uses the method of combination for target tracking.The method can play the role of inhibition when the target is seriously covered and is interfered by large area of color similar to the target,which tracks the target accurately.Finally, this text uses the improved tracking method to achieve the task of tracking target efficiently and accurately under the test results of binocular stereo vision method.By comparing with the traditional methods of detection and tracking, this article can effectively solve the interference of the similar color to target, and it can accurately realize tracking of single or multiple targets in the case of simple obscured with the binocular stereo vision inspection technology and the improved tracking method, which makes the precision more accurately.The final experimental results also show that the method can meet the target tracking task quickly and accurately.Besides, it also improves the precision of tracking greatly.