A new Voronoi diagram-based approach for matching multi-scale road networks

Abstract

AbstractObject matching is a key technology for map conflation, data updating, and data quality assessment. This article proposed a new Voronoi diagram-based approach for matching multi-scale road networks (VAMRN). Using this method, we first created Voronoi diagrams of the road network using the strategy of discretizing road lines into points and adding dense points to special road intersection segments. Then, we used the Voronoi diagram of road segment to find matching candidates. Finally, we obtained matching results by judging the geometric similarity metrics we designed and a heuristic combination optimization strategy. The experimental results demonstrated that the VAMRN outperformed two existing methods in generality and matching quality. The F-measures of VAMRN were 18.4, 29.6, 3.8, and 7.6% higher than the buffer growing method, and 4.5, 2.8, 1.8, and 6.1% higher than the probabilistic relaxation method. And the time performance is improved by more than 90% over the probabilistic relaxation method.