Research On The Detection Technology Of The Traffic Information Based On Video Image Processing

With the development of economy and technology, to establish one kind of all-directional, real-time, highly effective and accurate transportation information collection system is both inevitable and necessary. It becomes a research project in the transportation domain based on the video frequency technology volume of traffic collecting system because of its relatively small investment and a brilliant future. There are some problems of current mainstream video frequency imagery processing technique, not only for the ability of real time and anti-jamming technique is weak, but also have large amount of computation. This article mainly based on the video frequency imagery processing vehicles examination technology launches the work. Through the main problem in view of the present convention method to carry on improves and proposes some new examination method.Obtained a smaller, higher accuracy video image technology traffic flow magnitude examination system. Carries on the following research work:Firstly, in the detection the goal vehicles movement are followed by the shadow, the shadow's elimination is one of difficulty point in the video frequency imagery processing technique, its existence can create wrong examines. This article first obtains from the shadow elimination aspect, provide one kind of new multi-layered image fusion method to elimination shadow.Secondly, Vehicle tracking for itself with the characteristics, we can design a new method, through the vehicles movement, with the present frame vehicle speed and the acceleration forecast the future frame, and will unify the region the contour characteristic information to carry on the confirmation match, finally will accomplish the vehicles track.Thirdly, Design one kind of method to detect the vehicle speed, this method only need know that the vehicle in different time the position which located in the picture, no need to have large amount of computation, then may accurate figure out its reality distance,the distance divided by the time which the vehicle displacement spends then could obtain the vehicle speed.Fourthly, to achieve the traffic volume detection, design one kind of method examination belt technology the traffic flow magnitude video frequency examination., in which also uses the advanced version the background method of difference. Moreover the overall evaluation vehicles track the algorithm uses determines when the counting around whether is the identical vehicle.