Research And Implementation Of Moving Target Detection And Tracking Algorithm Based On OpenCV

With the rapid development of computer technology,the moving object detection and tracking has become a hot topic in computer vision.It combines artificial intelligence,image processing,pattern recognition,and morphological processing,and many other disciplines of advanced technology,the use of which related to technical input video image analysis and processing,so as to realize the target detection and tracking.The core of target tracking technology research includes the detection,extraction and tracking of targets.In this thesis,the difference between adjacent frames,three frame difference and background difference are studied,and the advantages and disadvantages of the target detection algorithm are compared and analyzed.In this thesis,the background weighted method is used to detect the moving target.Secondly,the basic principle of the traditional Meanshift algorithm and its application in video target tracking are studied.Through the experiment confirmed the Meanshift algorithm in the target occurrence of occlusion or complex background unable to track the target correctly,and in the target tracking process,tracking window can not be with the phenomenon of the change of target size.Finally,focus on the problem of tracking process in tracking window size can not be changed.In this thesis,in order to improve the bandwidth of the contour function,and to realize the tracking window adaptive changes,a new algorithm is used which is Based on the classical Meanshift tracking algorithm.To solve the occlusion problem for target tracking,this paper adopts the Kalman filter and Bhattacharyya combination method to the target occlusion problem to judge,according to the Kalman filter for moving target center coordinates of points and shape size prediction combined with Bhattacharyya coefficient of moving target is occluded to judge,Effectively solve the problem of occlusion.This thesis takes VS2010 as the software development platform,and designs the algorithm based on the OpenCV visual library.Finally,the results of the experiments are compared with the results of the test and tracking of the vehicle and pedestrian in the video.From the experimental results,we can see that the improved Meanshift algorithm has good real-time performance and anti-interference ability,and improves the accuracy of tracking of moving target.