Research And Implementation Of Moving Target Detection And Tracking In Intelligent Video Surveillance System

Intelligent video surveillance system has a wide range of applications in real life, it is through the analysis andprocessing of the video image to realize the detection and tracking of the moving object in the scene, and on th e basis of the analysis and processing of the target behavior, it gains people's attention and favor. This paper improves the accuracy and real-time of the intelligent video surveillance system in the target detection and tracking by the improved part of the m ainstream algorithm, and finally implements the function of the system based on the OpenCV. OpenCV is a goo d cross platform computer vision library, and it is a com pletely free product of open source, it can save a lo t of processing of developing before the preparation work, soit isconvenient to develop intelligent video surveillance system.The detection of the m oving target is the basis of the target tracking, and it is also the key to the analysis and processing of the intelligent video surveillance system. In this paper, we use the three fram e difference method for target detection, and combine with the image edge detection algorithm based on the im proved Canny operator to realize the tar get extraction. In the target tracking, this paper uses an improved CamShift algorithm to solve the problem of the error tracking which is caused by a large number of skin color regions in the image background. The experim ental results show that the system can achieve the function of target detection and tracking, and it also has good real-time performance and robustness.The work of this paper is mainly reflected in the following aspects:(1) This paper introducesthe reason of the inte lligent video surveillance system briefly, analyzesthe present situation and the developing trend of the intelligent video surveillance system, and expound the necessity of the res earch on the intelligen t video surveillance system.(2) This paper does som e research on the algorithm of tar get detection and tar get tracking in intelligent video sur veillance system, and im proved the f easibility and superiority of the algorithm on base of it, fina lly verified the feasibility and superiority of the algorithm by experiments.(3) This paper proposeda system framework of intelligent video surveillance syste m based on OpenCV,modularizedthe function of the system which was based on the requirement analysis and system flow, at last we designedand realizedthe system.(4) The system was combined with MFC, ST L standard template library and OpenCV and last we realized the whole function of intelligent video surveillance system.