New results with Generalized Area Partitionings

摘要

This paper presents the on-the-fly generalization resutls with the GAP-tree (Generalized Area Partitioning0 implementation applied to two large scale topographic data sets: vGBKN 1:1.000 and Top10vector 1:10.000. Applying the GAP-tree, it is now possible to interactively browse through these spatial data sets at medium 1:25.000 and even small <1:100.000 scales. In addition to the test resutls and implementation of standard CAP-tree, an extension is described based on adding parallel lines to the GAP-tree for linear features such as roads and waterways. This enables better, more smooth, enlargements of these features required at smaller scales. Further, a new method is given to build a GAP-tree representation for a very large data set based on a non-quadtree like divide-and-conquer technique.