Research On Traffic State Detection Method Based In Traffic Video

Ground transportation in order to meet the rapid development and resolve all traffic problems caused by traffic of a comprehensive analysis that the development of intelligent transportation system is referred to the important position. Remote vehicle detection and tracking system in the intelligent traffic management system is used as (ITS) in the important part, as the research focus of many research institutions. In this paper, the key ITS technology, research and ground-based video traffic in the vehicle tracking and detection of related issues.This article is in the video on the basis of vehicle detection, vehicle tracking results are discussed based on the traffic situation detection, analysis and research to determine the state of traffic conditions and proposed specific research methods, and the corresponding experimental process and results. Research conducted in this paper include the following main aspects: first, the video based on prior knowledge of the region of interest to be divided, in the vehicle motion detection algorithm, this paper, Gaussian mixture model to reconstruct the background to study the literature [35] in the Gaussian model-based shadow detection and removal algorithm, and to apply the algorithm to remove the shadow of this article, the experimental results more satisfactory results.After the detection of moving vehicles, the use of morphological processing images in the detection process to detect the noise occurs; again by the smallest rectangle marked on the test results. Movement of vehicles in the vehicle tracking algorithm, this paper based on kalman filter tracking algorithm with. The template in the value of each pixel assigned a Kalman filter to correct the pixel value changes, by a more gentle, reasonable template, based on the Kalman filter tracking and target illumination profile, occlusion have shown robust sex. On the basis of vehicle tracking, analysis and study of this road at some point the number of moving vehicles to determine the traffic conditions, such as congestion, Cheung Hang, and the traffic size. At the same time also proposed in this paper the specific detection method and experimental results. On the speed of detection, analysis and discussion of the calibration equation is derived and the possible cause of the error and so on.Described in this paper based on video tracking and detection of vehicle traffic to determine the status of proposed solutions have certain effects, and the experiment shows that this method in the analysis of traffic conditions is feasible, but results would be ideal.