Research Of Traffic Status Detection Algorithms In Its

Due to the rapid development of the economy, the number of the car in cities become more and more, which makes the road traffic congestion becoming frequently. The key to road traffic monitoring and management is to establish an intelligent traffic system. The correct detection of the traffic flow parameters and road traffic status discrimination is the key technology. This paper studies two modules of the traffic status detection algorithm:video-based traffic flow parameters detection and classification of traffic state based on classifiers.Video-based traffic flow parameters detection method has the advantages like having a low cost, easy to install, covering the large amount of information. There are mainly three methods to detect vehicle information based on video:the method of image frame difference, background subtraction and edge detection method. We analyze the simulation result of three algorithms on background extraction within traffic video. On this basis, use the algorithm of wavelet transform to extract component images of the background image. Compare to the three background extraction algorithms, use this method can effectively reduce trajectory of movement of vehicles in background images. Then use virtual-loop region to count the vehicle and detect the speed of it. The two virtual-loop regions are setting along the vertical direction of the road. The data in the virtual-loops region is processed in order to count the vehicle and detect the speed. The experimental results show that the method has certain feasibility.Also we propose an urban traffic state detection method based on support vector machine (SVM) and back propagation (BP). Fusing the SVM and BP classifiers into a cascade two-tier classifier improves the accuracy of the traffic state classification. We then fuse a cascade two-tier classifier SVM-BP, a single SVM classifier and a single BP classifier for the final decision so as to further improve the detection accuracy. The traffic data were collected from microwave data of Hangzhou Stadium Road to Fengqi Road (north to south) and from traffic simulation software VISSIM. We also propose fusion strategies at the training and implementation phases to improve the detection accuracy.The simulation tests show that the algorithms of this paper have a good detection and classification. The algorithms can be used as a reference to intelligent traffic system.