A quantitative similarity measure for maps

Abstract

In on-demand map generation, a base-map is modified to meet user requirements on scale, resolution, and other parameters. Since there are many ways of satisfying the requirement, we need a method of measuring the quality of the alternative maps. In this paper, we introduce a uniform framework for measuring the quality of generalized maps. The proposed Map Quality measure takes into account changes in all local objects (Shape Similarity), their neighborhoods (Location Similarity) and lastly across the entire map (Semantic Content Similarity). These three quality aspects measure the major generalization operators of simplification, relocation and selection, exaggeration and aggregation, collapse and typification. The three different aspects are combined using user-specified weights. Thus, the proposed framework supports the automatic choice of best alternative map according to preferences of the user or application. © 2006 Springer-Verlag Berlin Heidelberg.