| Recently,GIS technology has developed rapidly.From the early geographic information system with single function,simple structure oriented to professional department,it has developed into a comprehensive information system that integrates multi-function,intelligence,and network in all walks of life.The variety of objects determines the multi-semanticity of geographic information.At present,the study of geographic semantics focuses on the geographic ontology system based on the semantic concept description or the semantic model of the semantic attribute feature calculation,but it lacks the semantic analysis and recognition method of geographical objects that takes into account the expression of concepts and attribute.The study of semantic analysis needs to fully consider the attribute information contained in the semantics,the calculation of semantic attribute values and the relationship between semantics.This paper proposes a semantic analysis mechanism based on geometric algebra theory and method to construct semantic concepts hierarchy and attribute feature value fusion expression.The multi-vectors structure based on geometric algebra(GA)can realize the unified organization and expression of the mixed dimension geometric objects and semantic structure.The mathematical mode of geometric algebra supports the numerical calculation at the computer level,and geometric operator can effectively solve the semantic coding information,which achieved the semantic precise positioning and efficient identification.This paper researched on the general domain semantic library construction method based on domain ontology construction method,and defines the semantic library digital coding method based on the structure;the semantic coding and decoding expression method based on the mathematical formal structure and geometric operator of geometric algebra;set the weights of the semantic components and construct a semantic similarity measurement method based on the coding structure associated with GA concept description semantics and attribute constrained semantic.The main work of this paper is as follows:1)Domain semantic library construction:Learning from the "seven-step method" of constructing domain ontology,it acquires domain ontology knowledge and determines the concepts and the attributes of description concepts.Constructs a semantic database that accurately describes geographical entity structures and fully expresses semantic attribute information and its internal interactions.The database does not limit the type of storage objects,and supports the extension of attribute dimension and attribute space,which enriched the expression of semantics.Through the concept hierarchy of the semantic library and the classification of semantic attribute features,combined with the semantic coding rules to realize the transformation of semantic relations.2)Semantic information coding and decoding based on GA:Spatial base-vectors and geometric operators of GA can express specific semantic relations,The multi-vectors structure based on GA can realize the unified organization and fusion expression of semantic structure associated with conceptual objects and attribute data in mixed dimensions.Based on the semantic relationship of the original data,and constructed the semantic concept description,object association and attribute data embedding method based on GA to realize the integrity and consistency expression of the concept and attribute constraints of geographic objects.Based on the geometric inverse operator,the dimension information of multi-vectors can be extracted to realize the analysis of semantic information.3)Calculation of semantic similarity based on geometrical algebraic coding vectors:Based on the semantic coding vectors matrix structure of domain standard semantic database and study area semantic database,a semantic similarity measure model based concept hierarchy relations and attribute feature fusion coding structure is constructed to achieve similarity recognition between semantics. |
PDF Full Text Request