Research On Semantic Analysis And Computing Model For Geographic Instance

On earth, at least 60% of all information contains some geospatial reference. Geography elements are becoming important components of daily information one deals. Geographic information had been an important data source of a lot of domains, such as Environmental Protection, Smart City, Internet of Things, Location Based Service, etc. BUT it is so hard to extract meaningful geographic information from massive geographic data rapidly for non-geographic domain users, as geographic data resources continuously enrich. Moreover, there would be much more difficult to match increasing and various application demands by using geographic data only for a single user. Therefore, how to provide geographic information smart service has become an increasingly hot research area.Geographic information formalization is a suitable solution. In our study, semantic information extracted from geographic information was formalized as geographic knowledge and was served for geographic information computing. All of these processes, such as semantic information extraction, geographic computing, etc., cannot accomplished for non-geographic users independently for the reason of lack of professional background knowledge. The objective of this paper is to provide a case-solution for geographic data computing based on semantic information for non-geographic domain users. In this paper, a semantic-based geographic instance computing model (S-GIC) was proposed by introducing geographic knowledge, and transformed the model-center pattern to concept-center pattern.There are several innovative researches achieved as follows:(1) A geographic concept network based on concept-center pattern was constructed. The whole semantic information was divided into two parts, including semantic intension and semantic relationship among concepts. Semantic intension of concept was extracted and represented as ontological property set, and semantic relationship was formalized from geographic process models, whose output computing result was defined as the corresponding concept and input parameters were defined as the related concepts.(2 A geographic knowledge application ontology model-SGIC-Onto was constructed. The quintuple ontology structure was expressed as:SGIC-Onto={C, Pc, R, Ic, X}. Each element one by one in the structure referred to concept, ontology property set, relation set, concept instances and semantic constraints, respectively. In this paper, semantic analysis and formalization of geographic concept, concept semantic constraint and geographic instance data was performed. A core RDF/OWL vocabulary for geographic information based on the General Feature Model, Simple Features, Feature Geometry and SQL MM was imported to describe the spatial information of geographic instance.(3) The geographic concept network was designed followed by Linked Data four principles in order to realize knowledge sharing through internet web. All geographic knowledge, such as concept, instance, relationship, etc. was expressed as Resource Description Framework (RDF) triple form, including subject, predicate and object. And the access method of each resource was Uniform Resource Identifier (URI) style. Geographic knowledge linked database was constructed and published base D2RQ platform.(4) A semantic-based geographic instance computing model (S-GIC) was proposed based the geographic knowledge application ontology constructed above. There are two major components of S-GIC. The first one is the construction of geographic knowledge application ontology, including geographic concept semantic analysis, semantic constraints representation, etc. Another component is the technology implement of semantic-based geographic instance computing engine.(5) Before geographic instance computing, a data validity verification model based on ontology properties was employed. The model confirmed the semantic consistency between user input data and the need concept supported instance computing in terms of ontology property type. Finally, semantic accuracy was merged together with traditional syntactic one to ensure the correctness of computing results.(6) A prototype system based S-GIC model with some semantic operation functions, including concept semantic representation, semantic constraint construction, geographic knowledge chain computing, etc. was developed. And two experiments about instances extraction of Estuary and Slope of Slope was implemented, respectively. The results were proved some abilities of the S-GIC model, including semantic validity match, geographic concept network inference, etc. Finally, the case study on the urban heat island region extraction was implemented. And then it represented completely the geographic semantic information interoperation among each module of the S-GIC model and feasibility and rationality.