The Research On The Application Of Target Tracking Technology In Intelligent Video Surveillance System

Intelligent Video Surveillance System which based on Computer Vision technology has already been widely applied in various aspects of human’s production and life at present. Intelligent monitoring is integrated intelligent behavior recognition algorithms, it could be recognise and identify the behavior of the people or vehicle in the screen, and alert users when it’s in the appropriate conditions. Moving target detection and tracking is a very important intelligent technologies of video surveillance systems. This paper studies the intelligent video surveillance system implementation target tracking technology. The main tasks in this paper are as follows:(1) First, through study and research the traditional intelligent video surveillance system used three kinds of target tracking algorithm(Kalman filter algorithm, Mean shift algorithm, Adaptive mean shift algorithm); Then, carry simulation experiment of the three kinds of target tracking algorithms under different situations (no shelter, shelter half, full occlusion). Finally, analyze its advantages and the existing problems.(2) In view of the problem of poor occlusion robustness of the adaptive mean shift algorithm in the traditional intelligent video surveillance system when the target sheltered, an improved algorithm based on adaptive mean shift is proposed in this paper. Based on the Kalman filter algorithm has better predictive characteristics and adaptive mean shift algorithm has high real-time. This article is applied the adaptive Kalman filter algorithm into the mean shift algorithm, to solve the problem that when the target occurs occlusion the target tracking fail; it’s should be uses the Bhattacharyya coefficient to determine whether the target was covered during the tracking process.it can be predicted through Kalman filter algorithm when the covered occur,then the predicted result could be used to the next enter of Adaptive mean shift algorithm,the tracking target will be re-capture Quick once it’s appear. Simulation results of experiments to verify, the improved algorithm has a high real-time, and has strong robustness under the the situation when target occurs occlusion.(3) The system based on moving object detection and tracking is implemented on the Microsoft Visual Studio2008Integrated Development Environment Platform. The QT application program wizard and Intel OpenCV code are utilized to develop the system. The system can detect and track object in scene of input video from either USB camera or AVI files.