Research On Vehicle Detection And Tracking Based On The Large Trucks Illegal Lane Highway Monitoring System

With the rapid development of social economy and the increasing traffic volumes, the traditional way of traffic regulation is no longer able to meet the complex traffic situation, intelligent transportation systems came into being. This article comes from the subtopics Sichuan Science and Technology Office of Science and Technology Support Project "The major key technologies of intelligent traffic safety monitoring and flow control system".Project One:research on the key dynamic detecting technologies of the big trucks travelling in lines and the development of the monitoring systems. For this highway particular scene, researching the methods of vehicle’s detection, tracking and identification can effectively recognize and track the high speed vehicles, and identify large trucks accuratly, which provides strong support for establishing models of large trucks in illegal lane.Firstly, the article discusses kinds of theories invovling the collection of information and application scenarios of roads, and then compare the advantages and disadvantages of these methods and select a kind of method which can monitor intuitively and clearly, install expendiently, monitor a wide range of video surveillance. Analysis of target detections, tracking, basic vehicle recognition studies are displayed, various types of methods were introduced and being compared. Secondly, a new detection method is proposed based on space-time background difference detection model, which concludes the elimination of lighting and shadows, the filling of moving seeds,and the trace removing of moving target.This scheme makes it possible to identify the target becomes more clearly and complete. Finally, some experiments are applied aiming at the project proposed, further, the binary difference image comparison between the corresponding detection results and Gaussian mixture model is also displayed. The simulation result reveals the method can achieve a better rate of recognition, at the same time, the classification directing error samples and the error analysis are revealed. The final results prove that, in different intensity videos, using this detecting way can achieves a more than ninety percent accuracy.In addition, according to the characteristic of the same direction of vehicles running in highway, the article adopts the method of tracking which combines the vehicle moving position and color information, and the detailed principle is explained.Finally, by setting a threshold of the size of the vehicle through analyzing the nonlinear relationship in videos and detecting By selecting mass of samples in experiments related vehicle tracking, vehicle detection and identification, the article illustrates that this theory makes a foundation of establishing models which can identify the illegal lane in some technical study of testing driving trucks and the development of a control system.