Research Of Traffic Video Detection System Based On Image Processing

In monitoring of intelligent transport systems(ITS), traffic flow video detection is one ofthe important features in highway, bridge tunnels, urban roads and other places.Technicalconditions and task requests of the traffic flow video detection system are in constantdevelopment and change. Therefore,it is always focus and hot spot of intelligenttransportation systems research．Study and design about it also have a lage space fordevelopment.This paper mainly work is in image processing and movement in the inspection of thetarget.Use the current widely used in the field of consumer electronics open source imageprocessing libraries-OpenCV, enabling traffic detection.Main work contents are as follows:First of all, this article extensive researchs present situation of traffic video detectiontechnology, analysis video detection of technical difficulties, grasping the development trendof traffic video detection technology, clearly set out the technical difficulties and researchpriorities.Secondly, this article rounds image color space model, the gray color model, imageprocessing, image binary processing of mathematical morphology, image differenceoperation, conducted in-depth research. On the basis of the difficulties analysis in trafficvideo detection technology and primary processing programme, raised to the three inter-framedifference and background difference combined with the background of the operation toupdate technology, shadow detection based on color space conversion technology,characterized by virtual loop detection technology of vehicle statistics counting method.Finally,System program calls on OpenCV function-widely used in the field of consumerelectronics, provides the main image processing and motion detection code.Experimentcollected day and night traffic images.Experiments show that in the General section of the citytraffic day and night detection accuracy rate of up to100%.