The Matching Method Based On Voronoi Diagram For Multi-scale Polygonal Residential Areas

Matching multi-scale spatial entities such as residential areas is challenging because the positions, structure shapes and topological relationships of the same entities have significant difference. The traditional matching methods based on buffer or minimum boundary rectangles(MBR) usually cause missed matches or mismatch due to the deviation of entities' position in different scales. Furthermore, most of models of spatial entity similarity calculation are designed for the datasets with the specified map scale. Their applicability is too low to directly be used for the residential dataset with different map scale. Though, we present a general method of residential areas matching on multi-scale datasets based on Voronoi diagram in this paper. Using Voronoi diagram to automatically identify matching candidate sets instead of manually determined search space. Even when the sources datasets of matching contains large inconsistent position deviations, the method based on Voronoi diagram can get matching candidates effectively. Process of this method is as follows: first, to construct the Voronoi diagram for the small-scale dataset. Second, traverses each Voronoi diagram to find the appropriate large-scale features as the matching candidates. Finally, we use three similarity indexes to get the final matching results. They are convex-hull shape similarity, convex-hull area similarity and overlapping area ratio. We conducted experiments on the city of Zhejiang Province and Beijing. The scales of the tested datasets were 1:10000 and 1:50000, 1:1000 and 1:10000. The experiments showed that our method outperformed both the previous methods in generality and quality. Especially for the datasets where the inconsistent position deviations were large(i.e., the datasets of 1:1000 and 1:10000 scales), the average F-Measure of our results were 17.63% and 10.36% higher than the methods based on buffer and MBR.