Research Of Moving Objects Detection And Tracking Based On Video Surveillance

With the rapid development of the urbanization construction, the urban safety management has been not to be ignored, and Intelligent Video Surveillance System(IVSS) has been widely applied in the area of urban safety management. As one of the core technologies about software, the technology of moving objects detection and tracking in the video images has also drawn close attention. Thus, how to further improve the accuracy and reliability of the technology of moving objects detection and tracking in practice becomes the scholars’ research emphasis. This study analyses the main algorithm involved in the technology of moving objects detection and tracking and puts forwards suggestions about improving the feasibility of the technology through simulated experiment verification.Firstly, to improve the efficiency and effectiveness of the technology, this study describes the related methods about the image processing and analyzes the mutual conversion of image color spaces, noise reduction, image enhancement and mathematical morphology processing. Conversion of image color spaces and noise reduction are used in pre-processing technology, while mathematical morphology processing is used in post-processing techniques.Secondly, this study summarizes and contrasts the present popular methods of moving objects detection and tracking, and focuses on analyzing the Asymmetric Frame Difference Method and the Background Subtraction Method (BSM) in view of the Mixture of Gaussians Model (MoG). For the video images taken statically by single camera, the image setting is always fixed, but they could be influenced to a certain extend by illumination variations. This study presents the methods of moving objects detection and tracking mixing the Asymmetric Frame Difference Method and the Background Subtraction Method (BSM) based on the Mixture of Gaussians Model (MoG) to improve the robustness of illumination variations effect on detection and tracking. At the same time, in order to eliminate interference of shadow on detection and tracking, this study puts forwards a kind of shadow elimination method mixing HSV color space and vertical projection histogram aiming at exactly and efficiently detecting the moving objects even in the case of illumination abrupt variations.Finally, this study puts forwards tracking method of its own after analyzing the present frequently-used object tracking methods, which mixes the Mean Shift algorithm and Kalman filtering. With the intelligent fusion between the color characters and statistical characters in the target area as the matching basis, the method judges the area of moving object in the next frame to succeed in tracking. Meanwhile, this study also introduces the software design and realization of algorithm and verifies them by analyzing different videos.