Research Of The Key Video Processing Technology In Intelligent Traffic System

The rapid urbanization and transportation in today's society brings much pressure on traffic supervision. As an informatization mean, intelligent transportation system is capable of effectively protect the traffic safety, which can efficiently reduce manpower costs and increase the efficiency of supervision. So its research has attracted people's attention. The area and efficiency of an intelligent transportation system is one of the hot topics of research. In this paper, we mainly study the key video processing technology of intelligent traffic system for more efficiency and larger area supervision.In this paper, the intelligent transportation system mainly uses image processing technology to detect the vehicles. The fundamental of video image processing technique in the color space, grayscale, binary processing and so on is applied to process video data. As we also need to detect moving targets based on the algorithm of background difference, the multi-frame averaging algorithm to subtract the image background is also applied. At last, by using morphological processing method, the target image is optimized.Monitoring a larger area was bound to bring more complex image processing calculation of every frame. Therefore this paper launched a study about how to track the moving vehicles efficiently and exactly. A method combining the theory of median filtering and decimation was presented, which can reduce noise and reduce the amount of calculation of each frame image processing at the same time. Motion compensation algorithm based on the image's gray registration is used, which can effectively remove shaking noise. Considering the vehicle targets in the video, we used a target tracing algorithm based on inter-overlapping area after the achievement of the vehicle target detection, which can track the history positions of each target in the area exactly.In order to guarantee both the monitoring range and quality of illegal vehicles' capture, the capture methods are studied. We proposed a method called the prediction of the pixel coordinates of moving targets in the vedio. This method was convenient and practical, which was independent of any information of actual road or camera parameter. At the same time it can accurately predict the perpect capture position when capturing the moving target through zoom is needed. On these basises, zoom capture method in the intelligent transport systems was studied. And the least squares method was used to fit the road boundy, which can help get the monitoring road's datas. Then we can save the datas on disks for the reading of the datas next time. With the monitoring road's datas, we gave the critera and formulas of some illegal actions of the vehicles, such as crossing the illeagal line, reverse driving and emergency lane occupancy.Finally, this paper presented a construction method of a system called illegal vehicle intellegent tracking capture system. The system used some hardware devices like Dome Camera, industrial control computer and communication device. Aming to automaticly control dome camera to dynamically track the the moving vehicle target and analy their behavior, some software modules are used, such like dome camera controlling, image processing, intelligent judgment, illegal evidence and data upload.Tthe program was tested in an airport highway intersection, where the traffic accidents frequently occurred. It mainly got illegal evidence of vehicles that cross the illeagal line, reverse driving and occupy the emergency lane and achieved some satisfactory results. Therefore the program can effectively control vehicles' behavior and protect the traffic's safety.