Research On The Method Of Moving Target Detection And Tracking Based On OpenCV

Moving target detection and tracking is a new field of study.It combines control and digital image processing,pattern recognition,computer vision,artificial intelligence,image morphology and automation control,and many other fields of advanced technology.It is widely used in the visual warning,intelligent video surveillance,robot navigation and industrial testing video analysis and video processing.Based on the practical application as the background,focuses on the study of moving object detection and tracking method.Under the fixed background(camera fixed),detect moving targets and marked it in real time tracking,after in-depth research based on traditional detection and tracking algorithm.The experiment shows that the traditional method has some flaws.it is difficult to meet the requirements of different environments for detection and tracking on speed and accuracy.Especially in complex environments and multiple target environments.It is proved poor tracking results.In view of this,this paper do the followings:(1)In the part of the moving target detection,traditional detection algorithm is inter-frame difference method,background subtraction method and optical flow method.With the full analysis of comparative advantages and disadvantages to the various algorithms,it shows that traditional detection algorithm cannot achieve very satisfactory results under different environment.Therefore,based on the moving object detection algorithm with the traditional method,using the background subtraction and background modeling target detection algorithm for moving object detection.In different circumstances,realize the ideal of moving target detection.(2)In the part of the moving object tracking,nowadays,the most popular moving target tracking algorithm is Mean Shift algorithm,Kalman algorithm and particle filtering algorithm.Due to the background noise,clutter,occlusion and the fast-moving,classic tracking methods are not enough,sometimes result in error trace.In view of this,based on the particle filter tracking algorithm,improving proposed location--color histogram based particle filter tracking algorithm and improved algorithms at the same time.Finally,many experimental analysis of the effect of different algorithms,which verify the improved effectiveness and robustness of the algorithm.It is of great significance and practical value on modern intelligent video monitoring system development and application.