A Study On Moving Target Detecting And Tracking In Intelligent Video Surveillance Systems

Intelligent Video Surveillance technology is an emerging research orientation in the field of computer vision. Its main goal is to realize the description,understanding and analysis of the content of the surveillance video by integrating computer vision technology, digital image processing technology and pattern recognition technology. The analysis results can be used to control the video surveillance system itself. Thus, the video surveillance system will be improved to a higher level.The research contents of intelligent video surveillance include: moving object detection,object tracking, object recognition and object behavior analysis. Their research results can be applied into public safety protection, medical care, traffic management, customer service, and many other fields.Therefore, intelligent video surveillance has a profound theoretical value and broad application prospects.Moving object detection and tracking are two key technologies of intelligent video surveillance systems.This paper is commited to study these two key technologies:In moving object detection, first, we present a multi-modal foreground detection algorithm based on illumination invariant color features. Through deeply analysing the similarity matching method of the original algorithm, wo present a new similarity matching method under YUV space, and apply it to the background modeling process, and acquire a robust detection algorithm. Second, we have a study on a target detection algorithm based on codebook model, and present a periodic update idea to overcome the shortcomings of the codebook based detection algorithm, which can not adapt to the changes of background during the foreground detection process. So that the improved algorithm have some adaptive to the background variations. Then, we make a preliminary exploration on texture features based target detection algorithm, and achieved some good results.In the moving target tracking, we present a moving target tracking algorithm based on Kalman filter. First, foreground detection algorithms are used to detect foreground regions and extracted blobs, and then Kalman filter is used to build track, last of all, a nearest neighbor algorithm is used to find the correspondence between targets and blobs. Experiments show that, in the monitoring scenarios with sparse targets, the proposed target tracking algorithm can accurately track targets if foreground detection algorithm gives a good result.