Research On The Approach Of Road Traffic Information Extraction Based On Visual Technology

In order to solve the problems derived from the rapid development of urban traffic,Intelligent Transportation System(ITS)has become the focus of research at both home and abroad.Comprehensive,accurate and real-time traffic information provides the data support for the construction of ITS,which is the basis for traffic dispersion,road network planning and pedestrian route’s establishment,how to quickly and effectively extract traffic information is the key factor to restrict the development process of ITS.This paper will focuses on the research on the approach of road traffic information extraction.Compared with the traditional method of traffic information extraction,the traffic information extraction approach based on computer vision technology has become a hot topic in the field of intelligent transportation because of its convenient installation and maintenance,low cost and so on.However,because of the shadow of the vehicle and the ghost in the vehicle detection,the detection precision is reduced,most of the virtual coils used in the information extraction method need to be manually set and the parameters are difficult to determine,the application of traffic information extraction method based on visual technology still has some limitations.In this paper,the above defects will be studied,the main contents are as follows:(1)In order to solve the problem of the shadow exist in vehicle detection,a shadow elimination algorithm for vehicles in traffic video is proposed using the Principal Component Analysis(PCA).The algorithm has high robustness,no special requirement for traffic scene,no need of prior training and manual intervention;the introduction of PCA greatly reducing the computational complexity;compared with the traditional shadow elimination algorithm,the algorithm in this paper can improve the composite index of shadow elimination by more than 10%,and the computational efficiency is improved by more than 30%.(2)In order to solve the problem of the ghost exist in vehicle detection,a V-ViBe algorithm is proposed based on the ViBe algorithm with higher real-time performance.The algorithm constructs a "virtual" background image,changing the initial background model of traditional Vi Be algorithm,to inhibit the ghost at the source;using morphological knowledge to improve detection goal;the experimental results show that the proposed algorithm is superior to the original ViBe algorithm in terms of accuracy,recall rate,false detection rate,and other three performance indexes.(3)In the stage of information extraction,the lane line is extracted based on the prominent feature of lane line in the (4(7 space and the Hough transform principle;set up the virtual coil which suitable for lane shape according to the deformation coefficient of the lane line in the image;extract traffic monitoring video traffic information parameters based on vehicle detection algorithm in this paper.