On local heuristics to speed up polygon-polygon intersection tests

Abstract

The polygon-polygon intersection operation is CPU-intensive. Many data structures look into decomposing the polygons into multiple yet simple pieces to speed up the polygon-polygon intersection operation. This paper addresses local heuristics that can be adopted in these data structures by using local information about the simple polygon pieces to decide upon polygon-polygon intersections without having to perform this costly operation. The significance and effectiveness of each of the heuristics is studied. The paper also shows how these heuristics can be put together to perform a polygon join operation. Experiments are given to demonstrate the savings both in CPU and in I/O that result from these local heuristics.