Representing the Semantic Content of Topological Relations into Spatial Databases

An approach focused on incorporating semantic content into Spatial Databases is proposed. Our methodology is based on a conceptualization of a geospatial domain restricted to retrieve the meaning of topological relations between geographic objects by means of concepts. Indeed, in spatial databases only a small set of topological relations is explicitly represented and they are identified when the geospatial data are displayed or analyzed. While, a semantically enriched set of such relations may be required, sometimes this can be identified at the time when the geospatial data are manipulated by the user. Thus, we define six relations, which are obtained considering the behavior of diverse thematics such as Hydrology, Land Use, Transportation Networks, and Settlements. Geospatial objects are analyzed to identify the topological relationships. We consider two analysis levels: intrinsic and extrinsic. The intrinsic level consists of the analysis between geographic objects that belong to the same thematic. The extrinsic level is composed of relations between pairs of geographic objects that belong to different thematics. Therefore, descriptions are automatically generated in form of tuples {O i, R O j}, where O i and O j represent a pair of geospatial objects, and R represents the concept (relation). Each tuple represents the meaning of a topological relation. For example, a highway (O 1) crosses (R) a roadway (O 2). We consider that this method adds a partial semantic content to the geographic databases, because the concepts represent the meaning of topological relations. The conceptual representation has some advantages with respect to the traditional approaches: the conceptualization does not depend on the data scale, geo-reference system, dimension, etc. In addition, we propose a native format that has been designed to appropriately represent and analyze the topological relations.