Research On Detection And Recognition Of Newly Emerged Roads Based On Features Of The Vehicle Trajectory

With the development of the reform and opening-up,traffic projects for basic facilities in various regions have been changing with each passing day and provide significant support to the economic prosperity of China.Faced with the rapidly-changed and complicated traffic network,electronic map navigation such as AMAP and Baidu gives full scope to provide a good transport service.In addition,relevant government departments have invested great material resources and manpower in refreshing test of road network information.The mass trajectory data implies rich structural knowledge related to time,space and the traffic network.Therefore,this thesis,analyzing the correlation between features of the vehicle trajectory and newly-increased roads,advances the methods in regard to the recognition of construction regions based on the trajectory data of engineering vehicles and the detection of newly-increased roads based on that of taxis.The main achievements could be listed as follows.I.This research systematically investigated and surveyed the research status of the vehicle trajectory features and detection methods of newly-increased roads at home and abroad.On this basis,it preprocessed trajectory data,discussed the algorithm of vehicle infrastructure matching,the recognition of construction regions on account of the trajectory data of engineering vehicles and the detection algorithm of new roads based on that of taxis,as well as designed the system network and research methods of this study.II.With the mass trajectory data and road network information,this research adopted the inverted index principle to construct the mapping relation between grids and roads during the process of designing the algorithm of vehicle infrastructure matching,so as to improve the matching efficiency.As a result,it constructed the index mechanism to quickly obtain the tabulation of selected sections of highways from traffic trajectory data as well as achieved the precise matching between data and maps in multidimensional ways.III.Engineering vehicles,such as slag cars providing rich information of workplace and rate of progress through trajectory features,play significant roles in road construction.With the analysis of space-time trajectory characteristics of slag cars,this study devised the algorithm based on density clustering to achieve the recognition of construction regions.It set up the multidimensional optimized algorithm with a frequency of K and morning-evening comparison on actual operating rules of slag cars.The demonstration test between the experimental results and Map World in ArcGIS showed that the precision rate of construction region recognition could reach 92 percent.IV.On the basis of taxi trajectory characteristics,this research designed the detection algorithm of newly-increased roads on account of taxi trajectory data.The algorithm extracted unmatched track points,connecting them to form path information and adopting the method of path integration to generate newly-increased roads.It achieved the rapid and accurate detection through road attribute extraction.Finally,this study verified the detection results of newly-increased roads,with a randomly selected sample of 20 out of 183,in Xiapu County by Map World in ArcGIS,proving that the accuracy rate was up to 100 percent.The research findings of this thesis could offer useful reference to the renewal of electronic maps and the service of intelligent transportation.