Spatial reasoning using symbolic arrays

Abstract

Research in Artificial Intelligence and Cognitive Science has suggested the distinction between visual knowledge (such as shape, volume and colour of objects) and spatial knowledge (that is, spatial relationships among the different objects of a visual scene). We find this distinction applicable to Information Systems concerned with spatial reasoning and especially to Geographic Information Systems. In particular, this paper deals with the representation of spatial information in GIS. The paper presents a representational formalism which captures the knowledge embedded in spatial relationships and provides the ability to represent, retrieve and reason about spatial information not explicitly stored in memory.