Research And Application Of Traffic Flow Detection Algorithm Based On Video Processing

In order to alleviate the city traffic problem,Intelligent Transportation System(ITS),as an advanced traffic control technology,is attracting more and more people’s attention.Many cutting-edge technology about ITS has also been studied by a lot of researchers.The study of traffic flow statistics is the core content of the Traffic Monitoring System in Intelligent Transportation System.The main content of this paper is the car flow detection based on video processing.The traffic flow detection based on video processing is simple,fast and convenient,and it is superior to other detection methods in installation and maintenance.In this paper,the pretreatment of video,the extraction and tracking of vehicle targets,the calculation of vehicle count and the calculation of vehicle flow are analyzed in detail.In the vehicle detection algorithm,weighing the advantages and disadvantages of each method,the background subtraction method is chosen as the algorithm of vehicle detection.In the background updating,based on Statistic Histogram Method and Sliding Window Mean Filtering Method,this paper presents an automatic updating algorithm for a fixed window background.The method reduces the amount of data processing and can quickly update the background when the background changes in time.In the vehicle counting,combining the advantages of rectangular mark method and virtual line detection,a new method of vehicle centroid trip line detection is proposed.This method can detect the vehicle which cross the tripwire quickly.It also can reduce the rate of false.The traffic flow detection system developed in this paper is based on Visual Studio 2013 platform,using the OpenCV 2.4.13 codebase.The test results show that the vehicle flow detection system developed in this paper can achieve better detection results,and the accuracy is relatively high.