Research Of Moving Object’s Detection And Tracking Based On OpenCV

The technology of moving object detection and tracking, which is a hotspot in computervision, is combined with image processing, pattern recognition, artificial intelligence, and iswidely used in intelligent video surveillance, navigation and other fileds.Based on open source computer vision library and VS2010, this paper introduces criticalfunctions and common algorithms in OpenCV, like median filtering, gaussian smoothing, andoperator methods, laied foundation for later object detection and tracking algorithms thoughts.Through comparatively analysis of commom algorithms in recent years, a new algorithm isproposed. Firstly, this paper describes algorithms of moving object detection, mainly focuses oninter-frame difference, background subtraction and optical flow method. Based on the commonalgorithms, a new moving object detection algorithm is proposed. Secondly, through research oftracking algorithm, which is current popular, a new tracking algorithm is proposed in this paper.With the test of improved moving object detection and tracking algorithm, a detailed report isexplaned after each test result and the effectiveness and real-time of the improved algorithm isverified. So this paper has great significance for the development and application of intelligentvideo surveillance system.