Research On Automatical Detection Of Red Light Violation Based On Image Processing

With the development of modern transportation in our country and our city, especially the significant increase in the number of vehicles in recent years, the traffic management department has increased its investment on road safety and management year by year, which leads to more difficulty and complexity of the work. Among these, each traffic junction as part of authorities wins more and more attention by the traffic management department. In the management of each junction, the automatic identification and evidence of running the red light has become an important aspect of junction management. It has become more and more important to study how to upgrade or build a system of identification and evidence of running the red light under the condition of the existing low-cost and high-efficiency.This article has made an in-depth study of the key technology of the automatically detecting of vehicle running the red light, aiming at proposing an automatically detecting method of vehicle running the red light and filtering out the detecting effects of the background of changes, non-motor vehicles and pedestrians, at the same time, in preparation of checking and locating the vehicle number plates.In this article, through the different existing moving target detection algorithm analysis, and combining the environmental factors as running the red light, we propose vehicle movement detection algorithm based on the inter-frame difference. In order to eliminate the impact brought about by the movement of pedestrians and non-motor vehicles in the viewfinder box, we put forward the method of vehicle identification carried out through the car width and projected area of the body. Besides, in order to improve the accuracy and eliminate the fact that the remote objects become smaller in size because of the space perspective, we put forward the concept of projected area and the width of the proximal body and also give the calculation method.In this article, we test, in the detecting process of running the red light, the different situations that may exist at the crossroads, including going straight, turning right, and turning left, etc, and the traffic violations such as driving in the lane not according to the lane markers. In terms of the test, we propose the vehicle trajectory tracking algorithm, through the moving vehicles match, calculate the center of mass movement of vehicles, the tendency of change between the vehicle trajectory and the angle in the centerline, the tendency of change between the vehicle trajectory and the relative proximal distance in the centerline, then determine the line of vehicle movement, including left car turning left, left car turning right, left car going straight, right car turning left, right car turning right, right car going straight.In this article, through the analysis of various location algorithms of license plates, we put forward the location algorithm of vehicle number plate on the basis of the plate color, mathematical morphology and area matching.Finally, we devise an experimental system to carry out the test in terms of the vehicle detection algorithm, vehicle trajectory tracking algorithm and location algorithm of the vehicle number plate mentioned above. The result shows that the proposed algorithm can achieve such key technologies as the vehicle detection in the automatically detecting running the red lights, vehicle tracking as well as number plate positioning.