Traffic Parameter Video-based Detection Technology

With the development of digital image processing, video processing technology is gradually applied in traffic detection. Capture the prevailing situation of road vehicles by single or multiple cameras, analysis and process them with image technical, and then detect and identify vehicles, get the information of traffic flow,speed,roadway and so on. It provides important scientific basis for the macroeconomic road management and highway planning and design.In this paper, we use a period of video traffic images recorded by a stationary highway traffic cameras to be study object, do in-depth research and experimental to image Processing Technology, such as access and update of the background image,image segmentation and binarization, dilation and erosion, image denoising. Then analysis and design the video detection and tracking algorithm of vehicles based on the brightness curve. We mainly achieved vehicle identification in single lane and multi-lane,path tracking and production,detection and display of speed and traffic flow. Specific research contents and methods are as follows:(1)In the process of vehicles traveling, pixel brightness values of the vehicle region will get bigger (or smaller). Use this feature, drive average speed measurement based on image brightness curve is proposed in this article.First of all, interest regions are drawn by lanes, and generate the corresponding brightness curve of each interest region of interest generated within, and then according to the relationship of the brightness curve between adjacent frames, find the best matching region by minimizing error method and further estimate the speed. We can get the average speed of vehicles by doing weight average of a series of estimated speed.(2) Doing traffic statistics flow directly from the change of brightness curve.(3) Link the mobile trajectory of the center coordinates of vehicles, and the vehicles trajectory is generated. Clear out that path when the center coordinate is away from the video. The real-time trajectory of each vehicle is formed.(4) Design the interface, make every parameter display in the video in real time, and we can know the traffic situation of that time clearly and directly. This algorithm can effectively handle large amounts of data, and can effectively suppress noise, which can be applied to high-definition video monitoring. According to the experiment of the video captured in the real road, it shows the algorithm's superiority and reliabilit...