形态本体及其在地理空间数据发现中的应用研究

地理空间数据语义异构是实现数据关联、数据智能推荐和精确发现的主要瓶颈。地理空间数据本体被认为是解决地理空间数据语义异构的有效方法。形态特征是地理空间数据(除时空、要素内容外)的重要特征,是地理空间数据本体的重要研究内容。本文首先在系统分析地理空间数据形态特征的基础上,提出地理空间数据形态特征的概念体系。然后,建立地理空间数据形态本体模型,提出形态信息的本体表示方法,并构建地理空间数据形态本体。最后,基于形态本体的本体库,利用Jena本体推理技术,开发地理空间数据语义检索原型系统,并将形态本体应用于国家地球系统科学数据共享平台的元数据检索中。实验结果表明,地理空间数据形态本体可以有效地解决数据形态特征的语义异构,提高数据发现的查全率和查准率。本文的研究方法和成果对解决其他领域数据的语义异构,有重要的参考意义。

Research on Morphology-Ontology and Its Application in Geospatial Data Discovery

The semantic heterogeneity of geospatial data is the main bottleneck for the realization of data association, the intelligent recommendation and the accurate discovery of data. Geospatial data ontology is known as an effective approach to solve the semantic heterogeneity of geospatial data. The morphological characteristic is an important feature of geospatial data besides its temporal, spatial and attribute characteristics, and it is the primary research content of geospatial data ontology. Based on the systematic analysis on the morphological characteristics of geospatial data, this paper studies and puts forward a concept system. Furthermore, this paper creates the morphology-ontology model of geospatial data, defines the ontology representation method of morphological information, and ultimately constructs the morphology-ontology. In the last part, a prototype system for the semantic retrieval of geospatial data has been programmed by taking the morphology-ontology product as the ontology library and using the Jena Java framework. The constructed morphology-ontology have been applied to the retrieval of metadata from the National Earth System Science Data Sharing Infrastructure. Verification test shows that the morphology-ontology of geospatial data can effectively solve the semantic heterogeneity existing in the morphological characteristics of geospatial data and improve the precision and recall rate of data discovery result. The research methods and results of this paper have great reference values in solving the semantic heterogeneity problems occurred in other research fields.