Modeling and Formal Representation of Geospatial Knowledge for the Geospatial Semantic Web
Hong Huang,Jianya Gong
DOI: https://doi.org/10.1117/12.816008
2008-01-01
Abstract:GML can only achieve geospatial interoperation at syntactic level. However, it is necessary to resolve difference of spatial cognition in the first place in most occasions, so ontology was introduced to describe geospatial information and services. But it is obviously difficult and improper to let users to find, match and compose services, especially in some occasions there are complicated business logics. Currently, with the gradual introduction of Semantic Web technology (e.g., OWL, SWRL), the focus of the interoperation of geospatial information has shifted from syntactic level to Semantic and even automatic, intelligent level. In this way, Geospatial Semantic Web (GSM) can be put forward as an augmentation to the Semantic Web that additionally includes geospatial abstractions as well as related reasoning, representation and query mechanisms. To advance the implementation of GSM, we first attempt to construct the mechanism of modeling and formal representation of geospatial knowledge, which are also two mostly foundational phases in knowledge engineering (KE). Our attitude in this paper is quite pragmatical: we argue that geospatial context is a formal model of the discriminate environment characters of geospatial knowledge, and the derivation, understanding and using of geospatial knowledge are located in geospatial context. Therefore, first, we put forward a primitive hierarchy of geospatial knowledge referencing first order logic, formal ontologies, rules and GML. Second, a metamodel of geospatial context is proposed and we use the modeling methods and representation languages of formal ontologies to process geospatial context. Thirdly, we extend Web Process Service (WPS) to be compatible with local DLL for geoprocessing and possess inference capability based on OWL.