Traffic State Identification Method Research Based On License Plate Recognition Data

One of the main function of Intelligent Transportation System (ITS) is identifying road traffic condition, then provide real-time and reliable decision basis for traffic management and traffic traveler. Therefore, the traffic state classification and recognition is one of the highlights of intelligent transportation research. This article from the intersection of the license plate data of video detection equipment, studied the following three aspects of content:(1) Proposing the method of license plate fuzzy matching processing based on edit distance. The premise of researching traffic state discrimination based on license plate recognition data is upstream and downstream to obtain the license plate to do matching processing, although you can obtain more license data from a single intersection, but due to license plate recognition device error, upstream and downstream of license plate matching rate is not high, so based on the characteristics of the license plate information, put forward by calculating the edit distance of the finite time interval, deal with the license for fuzzy matching. Results show that this method increases the rate of license plate matching, improves the utilization rate of the license plate data and lays the foundation for follow-up study.(2) Realized the urban road traffic state identification method based on the license plate data. First of all, through the moment of two adjacent checkpoint to get the road bike travel time that the vehicle pass the road, and uses the distribution principle of wave and variation coefficient method to eliminate the abnormal bike travel time. Then, through optimizing the road bike trip time to obtain the sections of the interval average speed, using weighted historical data repair method to compensate the missing data, comparing with the traffic state discrimination standards it can realize the criterion of road traffic state. Finally, using a large number of the license plate data collected by the video monitoring system in Beijing to verify. The experimental results show that in this paper, eliminate the abnormal bike travel time out and missing data repair can judge the road traffic condition more accurately.(3) Implement the city intersection traffic state identification method based on the license plate data. Through the specific time of vehicle through the same intersection, obtain the time of intersection traffic lights change, combined with the specific time vehicles through the upstream and downstream intersection. Using the distribution principle of wave to analysis the corresponding phase within the same green time passing intersection vehicle queuing situation, judging whether the intersection of the vehicle has the second line, whether the intersection is in the saturated traffic state.