ABSTRACT
This chapter aims to provide an overview of research on the modelling aspects of spatial data handling. It focuses on the need for abstraction mechanisms which should generate high-level data descriptions based on prominent features. Descriptive modelling can be seen as the bottom-up approach to spatial data which has the main aim of upgrading knowledge of the observed natural phenomena by extracting information which is implicitly contained in the initial set of data. Descriptive modelling uses powerful interpretation algorithms which perform the analytical extraction of several features in order to generate a richer and more synthetic description, based on prominent features, which codes information in an efficient and effective way. Geometric models correspond to a second level of spatial data modelling and are derived from the basic model through the generation of relationships between individual observations. The geometric model reflects and highlights the particular nature of spatial data, which can be synthesised by two of its properties, continuity and coherence.
