Voronoi Feature Selection Model Considering Variable-Scale Map’s Balance and Legibility

Abstract

Variable-scale map because of its variability, destroys the constant of original scale, causing the map enlarge regional information easy to read, other regions are severely compressed and difficult to identify, reducing the map legibility. In this paper, we proposed a new pattern called Voronoi Feature Selection to solve the problem of information compression, considering map’s legibility and equilibrium. The main idea is that we use voronoi adjacency relationship model to select features, instead of traditional euclidean distance model, use voronoi influence ratio to determine the feature whether or not to remain, and remove the small influence features to reduce the loading of extrusion area, as well as improve the legibility of map. The comparative experiment results show that our methods make the transformed map readability and clearness to express, and it has a good feasibility.