Intelligent Monitoring System In The Study On The Algorithm Of Moving Target Detection And Tracking

Intelligent video surveillance has been a advanced and new topic of computer vision field. Withthe development of video technology and network technology, and now the video monitortechnology is moving towards intelligence, network, digital, and diversified developmentdirection.Video monitoring technology includes pattern recognition, automatic control,artificialintelligence, image processing, machine vision, and computer science and other advanced technology.The purpose of Intelligent video surveillance is to use the machine (cameras) to monitor and track themovement target insteading of people. Without the help of person,it is using computer visiontechnology, pattern recognition and digital image processing technology for automatic processing ofthe video camera. It can achieve target positioning, identification and tracking as a man. Comparedwith traditional monitoring system, Intelligent monitoring system has more advantages, it canautomatic identification, tracking and analysis; Reduce monitoring persons’ workload; Improvemonitoring efficiency; Reduce the chance of Misstatement. Now it is used with great success in thenational economy and social life and national defense and other fields, such as military visualguidance, robot vision navigation, safety monitoring, intelligent traffic, industrial product testing,medical image analysis, virtual reality, etc.The main contents of intelligent video surveillanceincludes following few aspectses：sports target detection, motion object classification, sports targettracking and understanding and description of the moving target. In the research of intelligent videosurveillance system, the most important technologies is The moving object detection and movingtarget tracking. Therefore, it’s of great practical significance and applicable value to research themoving object detection and moving target tracking.This article is to study the research of the moving targets detection and moving target tracking ofintelligent video surveillance system. First, it was to introduce the history of video monitoring and thedomestic and international present research situation. Second,the basic knowledge of imageprocessing is introduced, including the image color model, image filtering and image mathematicalmorphology. And then, it introduced several common target detection algorithm of moving objectdetection (background difference algorithm, frame difference algorithm., Optical flow estimation)and the classification of target detection. With an emphasis on the Prospects detection algorithm based on Gaussian modeling of background and modeling method of Cam Shift theory. Theadvantages and disadvantages have been analyzed and compared. With this understanding, this paperproposes a moving target detection and tracking algorithm based on frame difference and can-shiftalgorithm.On the research of moving target detection, on the basis of the research of three framedifference method, it proposes an improved three frame difference method used in the moving targetdetection algorithm, It can improve the rate of accuracy of the moving object detecting, with using themethod of combining calculation with sports template（the target image that what get by using finitedifference method）and contour detection. On the research of moving target tracking，in view of theCam shift algorithm (color histogram based on the mean translation algorithm) shortcomings, thispaper puts forward a sports target tracking algorithm which combination of a kalman filter and CamShift, using kalman filter forecast to test the position of the moving target and velocity size, andcombining with the Cam Shift, it can determine and track the target.According to the improved moving target detection and tracking algorithm, In order to validatethe performance of this algorithm, a program was established through the MFC developmentplatform and OpenCV function library. Experiments show that this algorithm can be a very goodway for the target of the detection and tracking in the video movement, it can achieve simpleintelligent judgment.