The Design And Realization Of Spatial Relation Reasoning Based On Ontology And SWRL

The formalization of spatial relation reasoning is a core part of spatialreasoning, a precondition of processing any spatial data, an absolutely necessary part ofspatial expert systems and geographic information systems (GIS), and an assistant toolfor spatial data mining. It has been widely used in site selection, way finding, mapmarking, diseases distribution and other activities that related to decision support.Originally, ontology is a theory that researches on entity existence and theessence of existence in western analytic philosophy. Afterwards, ontology wasintroduced into computer science domain as a knowledge representation tool. At present,as a tool for processing knowledge sharing, ontology is widely researched and used inAI, computational linguistics, GIS, database theory, semantic Web and so on.As respect to spatial relation reasoning, representing knowledge in spatial relationreasoning as ontology is not only favorable to the sharing and reusing of spatial relationknowledge, but also in favor of the reasoning. Especially, using ontology formalizationmodel which is based on description logic can fill up the deficiency in existingknowledge representations. But because present researches which are related to spatialrelation reasoning based on ontology are mostly focused on the representation ofontology in spatial relation reasoning, didn't pay so much attention to the research ofthe formalization of ontology and the spatial relation reasoning automation based onontology, so there isn't so many spatial relation reasoning method based on ontology.As a result, concerning spatial relation's ontology representation, this paper focused onspatial relation reasoning method research on the basis of analyzing spatial relationreasoning theory, ontology and related techniques. The core research content includes:①GIS is a main practical application domain of spatial relation reasoning. In orderto make SRC-Ontology——the generic core ontology in spatial relation reasoningusable in GIS, this paper expands and improves SRC-Ontology, so that we can use thisontology and the spatial relation reasoning based on this ontology in actual GIS to getspatial relation information from original geographic information.②Design the overall solution for spatial relation reasoning based on ontology andSWRL, this design has two parts, one is ontology reasoning, another is the reasoning based on SWRL. The ontology reasoning is aimed at making sure the establishedontology will have better structural levels, and simultaneously ensures there isn't anycontradictories in the definitions and conceptions which are generated from theontology; the reasoning based on SWRL is mainly used to realize the reasoning which isbased on the Combination of ontology knowledge and reasoning rules.③Based on the expanded and improved SRC-Ontology, use OWL DL to buildontology knowledge base in spatial relation reasoning, adopt Racer reasoning system,which is based on description logic, as the ontology knowledge's contradictories andconsistence detection tool to make sure the consistence and soundness of the ontologyknowledge. OWL DL is the sublanguage of OWL which is the recommended standardontology language of W3C, it is based on description logic. The ontology which isdescribed in OWL DL can not only bring about the sharing and reusing of knowledge,but also can use the reasoning mechanism of description logic to carry out ontologyreasoning fairly well.④On the basis of research on how to use SWRL to represent rules, this paperproposes the method of adding basic arithmetic in spatial relation reasoning to theSWRL language as user-defined BuiltIns by using the Java APIs which is provided byprotege. This method extends SWRL language, and then we can use SWRL to describereasoning rules in spatial relation reasoning.⑤In consideration of the situation that present reasoning systems can't processOWL knowledge and SWRL rules directly. This paper proposes to use SWRLJessTabto transform OWL knowledge and SWRL rules into Jess format, so that we can use Jessinference engine to carry out spatial relation reasoning.At last, based on methods proposed above, using Xi'An's tourism data as theexperimental data, this paper realizes a spatial relation reasoning prototype systembased on ontology and SWRL. The experiment indicated that as a spatial relationreasoning method, the spatial relation reasoning based on ontology and SWRL can notonly realize knowledge sharing and reusing, fill up the deficiency in present knowledgerepresentations, but also can realize rules sharing and reusing, and decrease spatialrelation reasoning system's complexity greatly.