Research On Moving Target Detection And Tracking Based On OpenCV

Since the development of our society and technology, video monitoring has been more important in security. Traditional video monitoring is rely on ma nual monitoring which can’t meet the need of social security today. That’s why smart video technology comes out. Smart Video Technology here, means analyzing video footage and monitoring,testing and positioning the objects in the area without the manual intervention. Monitoring, testing and positioning the objects in the footage is the key of this technology.(1) For objects monitoring, this paper introduced 3 methods for monitoring moving targets, including Optical flow method,inter- frame difference method and The background difference method.And figures out a new algorithm based on the background difference method. This algorithm depends on difference of the background to set the working mode. If the background changes slow, such as the outdoor situation, the algorithm will used the background difference method first, then will take a part of the background as the background update model. If the background changes fast, the algorithm will used the inter-frame difference method, then it will processed the images.(2) For tracking moving objects, after identifying the targets by the background difference method and the inter-frame difference method,the algorithm using the Camshift algorithm and the algorithm based on the Kalman prediction to tracking the specific targets. It will solve the problem which the Camshift needs the mouse to chose the moving targets or the problem which the targets may be covered or jammed by the targets with similar colors. The experiment shows that the algorithm can monitor and track the specific targets more efficiently.(3) At last, this paper test the algorithm by experiments based on the introduction of OpenCV, an open source Computer Vision Library system developed by Intel.