Research On The Key Technologies Of Intelligent Transportation Video Surveillance System

Intelligent video surveillance system whose aims at automatically analyzes image sequence to achieve target detection and tracking in the dynamic scene by using video analysis and alarming without people, is one of new arising high-tech application field and forefront topic of computer vision. Compared to traditional video surveillance system, it has advantages of higher quality of judging and less need of investment. It has cheerful prospect in the applications of surveillance for military, traffic, banking and other important monitor places.The dissertation aims at designing intelligent transportation surveillance system with static cameras. Based on studying the current conclusion, the dissertation analyzed, improved and realized the key technologies which relate to detection, tracing and recognition of moving object. The major contents of this dissertation can be summarized as follows:First of all, some common methods of moving objects detection were analyzed in the dissertation, and the method of background subtraction based on Gaussian mixture model was adopted. The dissertation proposed a novel algorithm which can effectively resolve the problems of background disturbance and light changes in allusion to the problem that the background subtraction is sensitive to light changes. The algorithm, by using the idea of simplifying background model, establishing of S and V component mixture model to reducing the computational complexity of the system during the course of background initialization and introducing the acceleration factor in the progress of background updating. At the same time, morphological filter and shadow removal based on HSV model was introduced in the progress of detection. Therefore, practicality and accuracy of moving objects detection system were improved.Secondly, Mean Shift algorithm was chosen to accomplish the task of object tracking. The dissertation improved the Mean Shift algorithm by predicting the target location in next frame of video with Kalman filter to solve the problem that the unsatisfactory effect of tracking fast moving objects with Mean Shift algorithm. Thus, the tracking speed and tracking precision was enhanced, and the reliability and robustness of object tracking system by the improved algorithm was improved too.Last but not the least, the paper discussed some methods of moving objects classification and adopted the Support Vector Machine for the classification. The function of the intelligent video surveillance is versatile, and the paper realized the application of moving objects detection, virtual cordon alarming, virtual area alarming and license plate recognition.The results of tests show that the algorithms in this dissertation are effective, which can be adopted in actual scenes to detect, tracking and recognition the moving objects.