Cross-section Traffic Flow Intelligent Detection System Based On Video Image Processing

With the rapid development of the current transportation industry,in the urban transportation network,there are often serious congestion at individual intersections,but the traffic at the nearby intersections is smooth.Therefore,real-time understanding of the traffic conditions at each intersection is important for the driver to choose a reasonable driving route,and is also an effective way to improve urban traffic congestion.The video processing based fleet length detection system combines image processing with various traffic information technologies.It has the advantages of wide application range,high measurement accuracy,good real-time performance and direct upgrade based on existing monitoring systems.It is a modern intelligent transportation.An important component of the technology of information acquisition.The process of identifying the traffic flow is to remove the noise of the image through the median filtering and Gaussian filtering of the vehicle video at the intersection,and then use the background difference method to detect the foreground vehicle and calculate the image by the B-spline curve method.The area of the vehicle so that the number of vehicles can be calculated.Among them,the median filtering module is improved in the 3×3 window median calculation method,and the method of sorting 9 data is used instead of the traditional calculation method,and the number of calculations is reduced from 30 times to 18 times,under the premise of ensuring accuracy.,saving machine time.In the background difference process,the background is first established,and the pixel points in the continuous image frame are subjected to differential operation.When the difference is smaller than the threshold,the point is regarded as the background point,and the background image suitable for the current situation can be obtained in real time through multiple trainings;The image is then subjected to a differential operation with the background image,and when the difference exceeds the threshold,the point is considered to be the foreground vehicle.By establishing a B-spline curve profile in the postdifferential image,the B-spline curve controls the normal direction of the point to approach the foreground target edge.After multiple iterations,the B-spline curve contour is the same as the foreground vehicle profile.The area covered by the spline contour is the vehicle area.Finally,the area is converted to the number of vehicles by preset parameters.The paper uses Matlab and Modelsim to simulate respectively.The test proves that the method can get accurate area and can accurately calculate the number of vehicles with an error of 10%.It has good detection accuracy and has high application in intelligent transportation industry.prospect.Figure [60] table [3] reference [73]...