An Adaptive Methodology for Automated Relief Generalization

Digital Elevation Models (DEM) are used in a wide range of applications. Several institutions worldwide are now involved in the collection of DEM data. OEMs are generally compiled at large scales (e.g. 1:25,000) with high accuracy. However, users frequently require data at different scales and for differing purposes (e.g. analysis, display etc.). To derive models at reduced scales, a generalization process has to be applied. This paper describes the development of an adaptive methodology for automated generalization of DEM data. The principal feature of this methodology is the adaptive selection of a suitable generalization method according to the scale of the resulting map and the characteristics of the given terrain: For smooth relief or minor scale reductions, a collection of filtering techniques (global and selective) is applied. For rougher terrain or larger scale reductions, a heuristic generalization procedure is used which works directly on the basis of structure lines.