Detecting Geometric Conflicts for Generalisation of Polygonal Maps

Abstract

Polygon generalisation is an essential process to produce multi-scale polygonal maps which are able to visualize categorical data at various levels of detail and with multiple themes. Detecting geometric conflicts, existing within an individual polygon and between polygons, is a primary step for polygon generalisation. This paper designs an inward–outward-buffering approach to detect narrow conflicts within an individual polygon. The approach not only detects the existence of narrow sections but also indicates their exact locations inside polygons. Another approach is presented to detect proximity conflicts in between polygons with two main steps. The first step associates buffering with an optimized proximity index to explore possible polygon clusters and spatial contexts around a problematic polygon. Based on the detected polygon clusters, the conflict regions are identified by combining outward–inward-buffering with topological operators. The tests with a real landuse dataset and two maps from other papers demonstrate that the presented approaches not only detect geometric conflict effectively, but also provide useful information for the selection of subsequent generalisation operators.