Multiple Paradigms for Automating Map Generalization: Geometry, Topology, Hierarchical Partioning and Local Triangulation

Generalization may be defined as a controlled reduction and simplifica tion of geographical data. Despite the knowledge of basic generalization operators, the automation of generalization remains a complex issue. Ac tually the change in resolution induces numerous spatial conflicts and the use of generalization operators such as smoothing or aggregation may ac cidentally degrade geographical data. The aim of this paper is to propose a set of paradigms necessary to automate generalization. Firstly a space partitioning computed from structuring objects allows to define some working areas and to give a pre-sequence of transformations in the proc ess of generalization. Then some local Delaunay triangulations allow to compute proximity relations between non-connected objects and to propa gate displacements due to some geometric transformations. To ensure consistency during different geometric changes, both paradigms are inte grated with a classical topo-metric structure.