The Research Of Video Image Target Tracking Algorithm

With the rapid development of mobile digital communication technology,multimedia technology and integrated circuit technology,the related applicationsof the video image information referring on social public service has becomemore and more popular in daily life.the construction of Video image processingtechnology as an important topic in the research and application, according president hujuntao advocating the theme of building a harmonious society,is more and more perfect in the field of security. At the same time, more and moreresearch,which is based on processing video image in recent years,is about intelligent monitoring technology,the video image information processing various technical indicators are also put forward higher request.In the field of computer digital information technology, the detection of video image target and the technology of processing image are a key technologys,the main research content of this article is the tracking and detection algorithm research of digital video image target and the movement target tracking algorithm,the tracking and detection algorithm research of digital video image target is based on computer vision algorithm processing in order to complish the video detection, target recognition and tracking, etc. the action of identifying target, Automatic detection and tracking target and recording can reduce the workload of video monitoring workers,by making the video image target detectionsystem intelligently,the monitoring system turns passive monitoring to active monitoring.the target tracking technology of Video image target tracking system is crucial and difficult, the research involves multidisciplinary field knowledge,computer image processing, pattern recognition,artificial intelligence and automatic control, such as government, academics and businesses realize the value of the subject research,it has high expectancy of application. Video image target tr acking technology requres forward higher performances on the stability and system accuracy, it is urgent to realize the adaption to the complex scene, high robustness, high accuracy of target tracking technology in current video image target detection.In this article, the Mean shift algorithm has carried on the detailed analysis, in view of the Mean shift algorithm in target under the condition of shadeeasy leakage problems, looking for a solution to this problem, after several tracking algorithm based on prediction estimates is analyzed and compared, the Mean shift algorithm and particle filtering algorithm is proposed combining method. Under the development environment of Opencv2.8.1,this article verify the algorithm by using a experimental traffic surveillance video. To solve the algorithm tracking precision rate has played a certain role in promoting.Target tracking system,based on video image in the key technologies including the choice of image features, target detection and target tracking and soon, has carried on the systemic analysis and research, the main work and innovation are as follows: using the color space of the HS two-dimensional vectorand scale invariant feature transform SIFT feature extraction of moving targets,combined with the Mean shift algorithm and the particle filter algorithm, which estimates target tracking algorithm improvement and verify accuracy of the algorithm, the effect of track estimates, the efficiency of the operation, etc.