Urban Road Extraction Based On Floating Car Trajectory Data

Urban roads are the main components of the urban transportation system,and they are also indispensable infrastructure for people’s lives.In order to alleviate the problems caused by rapid changes in the urban road structure and traffic congestion,it is necessary to extract and update urban road information in a timely and efficient manner.How to ensure the accuracy of the extracted roads and reflect the current road conditions as much as possible has become urban traffic management And research hotspots in the field of intelligent transportation.With the support of GPS positioning technology,the method of using wireless sensors to collect data has been rapidly developed and popularized.With the help of the floating car trajectory generated by it,road information can be directly extracted,compared with traditional surveying and mapping methods and remote sensing image extraction methods.,Obtaining road information is simpler,with short cycle,low cost,and wide coverage.Scholars at home and abroad have carried out certain research and results in this area,but the results of most urban road extraction at present are to extract road centerline or road width to express road information,and to express road intersections in the form of simple nodes,which cannot be accurate.To reflect the lane information inside the road,it is difficult to meet the increasingly accurate lane-level road network requirements of intelligent traffic and map navigation for urban traffic.In response to the rapid update of urban roads and the demand for lane-level road extraction in the field of intelligent transportation,this thesis studies the method of lanelevel urban road extraction based on floating car trajectory data,and provides technical and method support for obtaining more detailed urban road information.The main research of the thesis is as follows:(1)Lane-level urban road extraction based on trajectory direction.Aiming at the problem that the current urban road extraction results represent road information with the road centerline,this thesis studies the lane-level urban road extraction process based on floating car trajectory data,and proposes a clustering method based on the position and direction angle of the trajectory point.The method extracts the lane-level skeleton line from the clustered network of clustering points,and then constructs the smallest enclosing rectangle of the road polygon to obtain the road width information.(2)Obtaining road intersections taking into account lane information.Aiming at the problem of extracting the results of current road intersections with nodes,this thesis is based on the floating car trajectory data to obtain the relevant theoretical technology of road intersections.Based on the extraction of lane-level urban roads,the trajectory points between road segments are taken as the research object.Identify intersections through road staggered relationships and defined range circles,and generate trajectory lines,trajectory clustering,and trajectory line fusion by connecting track points according to the direction angle change of the floating car trajectory at the road intersection and the neighboring relationship between the trajectories Steps to obtain as comprehensive and accurate road intersection information as possible.(3)Experimental analysis of urban road extraction.Based on the extraction method of urban roads and road intersections proposed in this thesis,experimental analysis is carried out using Chengdu taxi trajectory data as the data source to verify the lane-level urban road extraction method proposed in this thesis and the road intersection acquisition method that takes into account lane information.Through experiments The analysis shows the rationality and effectiveness of this method.