Research On Vehicle Travel Feature Based On License Plate Recognition Data

With the development of social economy and the improvement of people’s living standards,the number of private cars is increasing.In some small and medium-sized cities,as the public transportation network is not yet perfect,private car becomes travel mode of normal commuters and others because of their convenient and fast advantages.However,with the acceleration of urbanization,urban congestion problems,frequent traffic accidents,serious environmental pollution,inadequate parking facilities,and outstanding dynamic and static traffic problems have become prominent.In response to such traffic problems,the traffic management department can rationally guide the traffic travel demand by formulating relevant traffic demand management policies,thereby alleviating traffic congestion.It is of great practical significance to classify and analyze the vehicles in the road network,determine the traffic flow structure in the road network,and clarify the usage characteristics and travel rules of various vehicles for formulating targeted traffic demand management policies.Public transport vehicles,taxis and commuter vehicles have high frequency of use and large traffic impact.It is crucial to accurately distinguish these three types of vehicles,and study the temporal and spatial distribution of traffic travel.The traffic data collected by intelligent traffic devices such as HD bayonet and electronic police equipment provides important support for studying the vehicle’s use characteristics and travel rules.Based on the license plate identification data,this paper studies the characteristics recognition of vehicle travel.The main contents are as follows:(1)In terms of the quality of license plate recognition data,this paper discusses the data quality of license plate recognition data from two aspects: flow accuracy and license plate recognition accuracy,combined with video traffic data and manual survey flow data.(2)Considering the data redundancy,data loss,data error and data anomaly in the original license plate data,the identification method and pretreatment process are proposed,and the case analysis is carried out in combination with the data of a certain city to explore the proportion and influence of various abnormal data.(3)Using the method of statistical analysis,the traffic characteristics of urban road network are studied from the two dimensions of traffic travel characteristics and vehicle use characteristics,and the traffic phenomena such as urban travel demand,traffic travel structure and traffic travel time are analyzed.Vehicle use characteristics are analyzed from three dimensions: vehicle usage degree,vehicle attribute correlation and vehicle use strength stability to explore the use of urban vehicles and analyze the distribution of vehicles in different attributions and the travel distribution of vehicles belonging to different attributions in time.Finally obtaining main body of urban traffic.(4)Filtering and classifying data.All the data of the working day are extracted and preprocessed,and the pre-processed data is classified according to the travel characteristics of the three types of vehicles: buses,taxis and commuter cars,and the unrelated vehicles that have influences on the cluster are eliminated,and three types basic data set of vehicle analysis are obtained.According to the travel characteristics of the three types of vehicles,respectively extracting the feature indicators to obtain the index set of cluster analysis.(5)Taking the actual license plate data as a sample for case analysis,the three types of vehicles,such as bus,taxis and commuter,are extracted,and the travel characteristics of the three types of vehicles and the influence of the three types of vehicles on the road network are analyzed.The clustering results are verified to illustrate rationality and effectiveness of the clustering method and index selection.