Research On Voronoi-diagrams-based Matching Method Of Multi-scale Road Networks

In recent years,with the rapid development of mobile Internet and GPS technology and the rise of VGI(Volunteered Geographic Information),the public has the freedom to edit and create maps on the open map platform,and voluntarily publish the contents by themselves on the Internet to provide to other users,thus crowdsourcing geospatial data suddenly come into our sight with large amounts of data,rich in content,real time,and free access.Crowdsourcing geospatial data are widely used in data integration and updates,stop-losses and rescue after disaster,LBS(Location Based Service)and many other fields.However,due to non-professional and non-standard production,crowdsourcing geospatial data are greatly different from traditional professional surveying data in geometric,semantic,and topology,how to quickly and effectively integrate professional mapping data and crowdsourcing geospatial data becomes a huge challenge for GIS(Geographic Information Science).With the accelerating pace of urban road construction,the demand for up-to-date road network information in the fields of navigation,intelligent city and intelligent transportation is increasing.The entity matching is the basic way of road map conflation and updates.Hence,it is of great significance not only in theoretical research but also in practical applications to study road matching.This paper takes road datasets as the experimental objects to conduct research on multi-scale road network matching,so as to solve the problems of low matching accuracy,dependency on radius buffer and poor genarality.The main results of the work are as follows:(1)The research background and current situation of road network matching were analyzed,and the present problems were pointed out.This paper introduced the basic concepts and processes of entity matching,as well as the commonly used similarity indexes and the quality evaluation methods of matching results.(2)The construction method of road Voronoi diagram was studied.The ordinary Voronoi diagrams and the generalized Voronoi diagrams were introduced.According to the generalized Voronoi diagram theory,the Voronoi diagram algorithm of the road network was studied.(3)The similarity evaluation indexes of road network matching were designed.The differences of the similar scale and trans-scale identical roads were analyzed from the point of view of human cognition,based on which the similarity indexes of length,shape,and topology were selected and designed for similar scale road network matching,while shape similarity index and area of the closed curve similarity index were for trans-scale road network matching.(4)The Voronoi-diagrams-based matching framework and strategy of multi-scale road network were proposed.First,the initial candidate matching set was filtered based on the Voronoi diagrams,and a combination matching strategy based on directed Hausdorff distance and a combination matching strategy with neighborhood similarity were then put forword to construct the Voronoi-diagrams-based matching method of multi-scale road network.(5)With several multi-scale road network datasets,the experimental platform was designed and implemented by Microsoft Visual Studio 2013 and ESRI ArcGIS Engine 10.2.The experimental results showed that the proposed method was feasible and had strong universality,and it could not only identified the matching relation of 1:1,but also could well identified the 1:N and M:N matching relations,When there was a large non-uniform deviation with the matching data,it could also acquired high recall and precision.