Research On Road Extraction Methods Based On Trajectory Data

As the foundation of geographic information system,high-definition map is significant to intelligent city,vehicle navigation and autonomous vehicles.With the development of social economy and the increasing of urbanization,road networks in urban areas are also changing frequently.However,traditional mapping methods which are costly and time consuming cannot reflect the road changes of city road network in time.Updating road network with vehicle trajectory data is now a hot topic among many researchers.The development of navigation technology and wireless communication technology has led to an explosive growth of trajectory data.these massive trajectory data cost little,update frequently and contain large amounts of information,making it a good source for road updating.This paper is based on trajectory data and the main work is as follows:(1)The raw GPS data collected from taxi is analyzed and a storage model is designed.(2)A data clean algorithm is proposed to reduce the redundancy of trajectory data.(3)A distributed parallel algorithm is proposed to accelerate spatial buffer analysis.Road change information is extracted based on this algorithm.(4)Road change information is clustered by a density-based method,and road centerlines are extracted from the clusters.Then an algorithm is proposed to build topology between extracted road network and origin road network.