Pedestrian abnormal Intelligent Video Surveillance System Research and Implementation
|School||University of Electronic Science and Technology|
|Course||Computer Software and Theory|
|Keywords||Intelligent Video Surveillance System Target tracking Abnormal behavior detection Moving object detection|
Intelligent video surveillance technology is an important direction in the field of computer vision, but also the research topic of concern, involving knowledge of computer vision, image processing, artificial intelligence, pattern recognition, and other disciplines. Intelligent Video Surveillance System with traditional monitoring system maximum difference is that with intelligence, through the automatic analysis of surveillance video of the content, to achieve the monitoring scene movement target detection, identification and tracking, and ultimately target behavior detection, given to the camera the functions of the human eye. Intelligent video surveillance system has been in the Beijing Olympic Games, Shanghai World Expo and other major occasions shine, with the emphasis on security and surveillance technology mature, intelligent video surveillance system will go deep into every corner of society. This paper combined with the latest research theory of computer vision, learning using computer visual development platform OpenCV, get pedestrian tracking the trajectory carried abnormal behavior detection, the final formation of the able monocular fixed camera head carried pedestrian abnormal behavior detection of intelligent video monitoring system. The main contents are as follows: 1, moving object detection, moving object detection method based on Gaussian mixture background modeling. This method not only applicable to the detection of a simple scene, the monitoring scene such as: rain, snow, leaves a slight swing complex, testing the effect is very good. 2, target tracking, the basic algorithm: Kalman filter algorithm, the mean shift algorithm, particle filter algorithm in-depth study and compare the advantages and disadvantages of the three methods on the object tracking. Comprehensive Kalman filter algorithm calculation speed, the mean shift the characteristics of high accuracy and fast convergence of the algorithm and particle filter algorithm is proposed based the clumps adaptive object tracking method. The algorithm briquettes state is divided into three cases: no occlusion, partial occlusion and severe occlusion, the choice of the type according to the different states can be adaptively tracking algorithm. In the the unobstructed case according to the distance or similarity of the briquettes briquettes match. In the case of clumps of partial occlusion, the mean shift tracking algorithm. In the case of the clumps serious blocked, using the particle filter tracking method. Experiments show that this method can be a good solution to the clumps collide when moving object tracking problem. 3, the detection of the abnormal behavior, wandering through the analysis, falls, the three characteristics of bounds behavior, based on the trajectory of the pedestrians wandering pedestrians crossed the pedestrian falls the three abnormal behavior detection method, according to the results of the moving object tracking. 4, the the mass tracking framework OpenCV based integrated moving object detection, target tracking, behavior analysis and other important functions of the intelligent monitoring system module, designed and implemented a complete pedestrian abnormal intelligent video surveillance system. The system is based clumps adaptive tracking algorithm to extract the trajectory of the moving object by analyzing the trajectory of the pedestrian, the pedestrian falls, wandering pedestrians, pedestrian cross-border the three abnormal behavior detection. The experiment proved that this system is capable of real-time processing of video data, and is able to detect the pedestrian falls, wandering pedestrians, pedestrian cross-border three abnormal behavior.