GIS空间知识的形式化表达与集成方法研究

发布时间：2018-02-27 06:33

本文关键词： 空间知识地理本体空间知识形式化表达空间知识集成　出处：《南京师范大学》2013年博士论文　论文类型：学位论文

【摘要】：在GIS学科中,空间知识是指在不同认知层次上一个或多个地理信息关联在一起形成的有应用价值的信息结构。针对空间知识的研究是GIS智能化方向的核心内容,也是空间数据挖掘、地理空间推理等课题的基础,多侧重于通过空间数据挖掘等方法实现隐式空间知识的获取,以及在实际生产中应用空间知识进行辅助决策分析,提高问题求解能力。当前,多源异构空间数据的集成技术已经相对成熟,而空间知识仍缺乏通用的形式化表达方法,不同来源、不同领域、不同方法获取的空间知识普遍存在语义、空间尺度、表达方法和存储结构等方面的异构性问题,导致空间知识的共享和集成困难,制约了通过跨领域空间数据和空间知识进行综合性应用的智能化水平。另一方面,由于传统知识工程领域侧重企业业务性知识集成的研究成果尚无法直接适用于空间知识集成这一命题,空间知识的集成还缺乏理论模型及方法的支撑。因此,研究空间知识的统一形式化表达与集成方法,消除空间知识的异构性,顾及空间知识的时空特征,从而可以有效利用当前多源异构的空间数据与空间知识资源,为今后在大数据量环境下的空间数据挖掘提供新的思路,丰富空间知识研究的理论方法体系,并对当前的智慧城市建设起到一定的方法性指导作用。本文以地理信息的多源异构空间知识集成方法为主线,以构建空间知识的统一表达模型为切入点,在梳理、明确空间知识的内涵、分类体系与特征因子的基础上,针对当前空间知识存在异构性和知识集成方法对空间知识中时空特征处理能力不足的问题,通过改进的地理本体提出一种空间知识的统一形式化表达模型,实现对空间知识特征因子的完备表达,并通过转换、合并及推理实现空间知识的集成。本文取得了以下的研究成果:(1)提出一种空间知识的统一形式化表达模型(Geospatial Knowledge UnifiedRepresentation Model,简称GeoKURM)。对比典型的空间知识表达方法,分析其差异性,证实地理本体最适合空间知识的统一表达与异构表达方法的集成。归纳构成空间知识形式化表达的特征因子,通过空间元知识、空间知识本体以及空间规则库实现对特征因子的完备描述。提出一种改进的语义三元组来描述语义因子;针对地理本体中拓扑关系表达能力的不足,基于4交集模型提出的GeoKURM的拓扑关系描述,支持一致性推理和冲突消解。(2)提出一种空间知识的集成方法,通过表达方法转换、语义映射与合并、以及推理检验三个步骤实现。在空间知识的转换方法中,提出空间知识转换的一般流程,对空间描述逻辑建立拓扑关系描述的映射关系以实现转换;对结构化表达的产生式规则提出通过时空关系词汇集实现转换时的隐式时空关系自动识别与抽取方法,解决空间知识的表达方法异构及存储结构异构。在空间知识的合并方法中,基于本体映射方法进行GeoKURM的合并,解决空间知识的语义异构。在空间知识推理中,基于GeoKURM的时空特征表达提出一种改进的一致性检测推理机GeoPellet,与Jess规则推理机构成GeoKURM的混合推理机制。(3)设计空间知识处理平台GeoKMP及原型系统实现,支持领域空间知识库的构建,异构空间知识源的集成以及结合实际数据的综合推理。以城市综合管线的数据校验和城市燃气管线的危险源辨识为应用案例,集成多领域空间知识库并整合相关空间数据库资源,优化改善已有的空间知识库,验证本文提出方法的准确性。
[Abstract]:In GIS subjects, spatial knowledge refers to the information structure in different cognitive levels of one or more geographic information associated with application value together to form. According to the research of spatial knowledge is the core content of GIS intelligent direction, and spatial data mining, spatial reasoning task, focus on the spatial data mining method to obtain implicit knowledge of space, and the application of spatial knowledge in actual production decision-making analysis, improve the ability of solving problems. At present, the integrated technology of multi-source and heterogeneous spatial data has been mature, but still lack a general spatial knowledge representation method, different sources, different areas of space the knowledge acquired from different methods are common semantic, spatial scale, heterogeneous expression method and storage structure, resulting in the spatial knowledge sharing and integration difficulties, system The intelligent level about the comprehensive application through the cross domain of spatial data and spatial knowledge. On the other hand, because of the traditional knowledge engineering business enterprise knowledge integration research achievements can not be directly applicable to the space of knowledge integration this proposition, integration of spatial knowledge is the lack of the theory model and method of support. Therefore, with the integrated method of expression of the unified form of research on spatial knowledge, to eliminate the heterogeneity of spatial knowledge, take into account the temporal and spatial characteristics of spatial knowledge, which can effectively utilize the space data of the multi-source heterogeneous spatial knowledge and resources, to provide new ideas for the future of spatial data in a large data mining environment, enrich the theoretical system of the spatial knowledge, and to a certain method of guidance on the wisdom of the city. The construction of the multi-source heterogeneous spatial knowledge integration based on geographic information. As the main method, the expression model as the starting point, to unify the construction of spatial knowledge in combing, clear connotation of spatial knowledge, basic classification system and characteristic factor, according to the existing heterogeneous and knowledge integration method of spatial and temporal characteristics of processing spatial knowledge in the problem of spatial knowledge, proposes a unified expression model the form of spatial knowledge through improved geographic ontology, realize the complete expression of the spatial knowledge characteristic factor, and through the transformation, integration of spatial knowledge merging and reasoning. This paper has achieved the following results: (1) the expression model presents a unified form of spatial knowledge (Geospatial Knowledge UnifiedRepresentation Model, referred to as GeoKURM). In contrast to the classical spatial knowledge expression method, analysis of their differences, that unified geographic ontology most suitable spatial knowledge expression and heterogeneous expression The integrated method. Summarized constitute characteristic factor expression of space knowledge, meta knowledge through space and space, space ontology rules to achieve a complete description of the characteristic factors. This paper proposes an improved semantic three tuple to describe the semantic factor; for lack of geographic ontology in the topological relationship between the ability of expression, to describe the topological relationship between the 4 the intersection model is put forward based on GeoKURM, support the consistency reasoning and conflict resolution. (2) proposed a method of integration of spatial knowledge, through the expression conversion method, semantic mapping and merging, and reasoning test three steps. In the spatial knowledge conversion method, put forward the general process of knowledge transformation of space, space the description logic mapping describing topological relations in order to achieve the conversion of structured; expression of generative rules proposed by the temporal and spatial relationship between the vocabulary set implicit conversion when the time The relationship between the automatic identification and extraction method, solve the expression method of spatial knowledge and heterogeneous storage structure. In the combined method of spatial heterogeneous knowledge, GeoKURM with ontology mapping method based on spatial knowledge. To solve the problem of semantic heterogeneity in spatial knowledge inference, spatial and temporal characteristics of the expression of GeoKURM presents a consistency detection reasoning machine GeoPellet improvement based on the hybrid reasoning mechanism is composed of GeoKURM and Jess rules of inference. (3) design and implementation of spatial knowledge processing platform and GeoKMP prototype system, build the support field of spatial knowledge database, heterogeneous spatial knowledge source integration and combined with the actual data. By reasoning city comprehensive pipeline data check and city gas pipeline risk source identification for application case, space integrated multi domain knowledge base and integration of spatial database resources, optimizing the spatial knowledge base to improve the existing inspection. This article suggests the accuracy of the method.

【学位授予单位】：南京师范大学
【学位级别】：博士
【学位授予年份】：2013
【分类号】：P208

【参考文献】