| Abstract | On the basis of user specifications (map scale, map purpose, symbology, and information density) the appropriate object classes are selected from the LoD for the desired map scale and further refined at nun-time by on-the-fly generalization algorithm. This paper focuses on the discussion of the former issue, trying to develop cartographically sound generalization processes that exploit the combined strength of new approach the interpolation of spatial objects for implementing with generalization and multi scale dataset. Tile central idea is to combine two used approaches in cartography: on the one hand the use of a multi-scale datasets which includes two or more levels of details, on the other hand the use of new approach for the interpolation of spatial objects. For the approach of this paper is based on the interpolations. One of the most popular fields is image interpolation. The image interpolation is a general term for a set of techniques used in image synthesis to generate intermediary pictures between two successive images. The interpolations performed by using mathematical morphology iransformations meaning, morphological methods are based on mathematical morphology, which deals with different types of images e.g. binary, mosaic, graytone and color. The interpolations are-based on a transformation, such as geodesic operators or ha usdorff distance computations can be used. "There are various ways to define interpolation between two sets of a metric space. If' two datasets are supposed to be compact and non empty, hausdorff distance provides one possible approach. The hausdorff metric admits at least two different geodesic, with provide interpolators between pairs of sets. When metric space is aftine, the space quotient under translation admits in turns reduced geodesics, which generate better set interpolators. Finally, by introducing a hausdorff distance by erosion, one can provide a symmetric form to interpolator, but are no linger geodesics. The median element and hausdorff distances are applied with vector objects. An approach for spatial objects interpolation is developed and implemented. This approach can manipulate both; convex and concave without convex decomposition in one time to interpolation the spatial objects |