EDGECRACK: a parallel Divide-and-Conquer algorithm for building a topological data structure

In this paper we consider the problem of converting a large database of 2D polygons into a topological data structure (a data structure with nodes, edges and faces). We present EDGECRACK, an algorithm to obtain the topological data structure (which is based on a known algorithm for segment intersection) and performs small geometric corrections of the input by snapping to avoid problems. We further show how we have extended this algorithm to a Divide-and-Conquer approach, which is also suited for parallel processing. We present experimental results based on our implementation and show that we have been able to convert a large database of 5.3 millions polygons into a topological data structure.