Deriving Multiple Representation Database: A Model Generalisation Approach

Abstract. The main aim of the research being conducted is to automatically derive multiple representation database through model generalisation. In this regard, this paper presents a detailed methodology of the derivation process of multiple representation database. The topographic information of three different levels of detail (LoDs) viz., LoD1 (1 : 10,000), LoD2 (1 : 25,000) and LoD3 (1 : 250,000) are considered for this research. The abstract LoDs, i.e., LoD2 and LoD3 are derived automatically by model generalisation. The method for linking corresponding features in all these LoDs is shown and thus deriving multiple representation database. The major component of the proposed methodology is the model generalisation process. The design of the model generalisation process which includes generalisation operators, classification hierarchies, geometric constraints, functional hierarchies and topological constraints are discussed in this article. The proposed methodology is currently in the implementation stage.