Research On Geospatial Data Provenance Sharing Based On Linked Data Approach

Geospatial information resources as the important component of information resources, is widely used in disaster monitoring, environmental assessment, military operations, the urban and rural planning, transportation, agriculture, forestry, meteorology and other fields. With the rapid development of geospatial information technology and earth observation technology, the spatial data product is growing exponentially in web and user's demand for spatial data is increasing as well. The processing and integration of spatial data has become increasingly frequent. How to obtain spatial data production process information, judge the reliability and credibility of spatial data, and identify the quality of data, which had become one of important problem in the field of geospatial research. Spatial data provenance can resolve this problem. In web services environment, the tracking, express, storage, query and sharing of the geospatial data provenance has become a new research hotspot in the field of spatial information.The existent research of geospatial data provenance have some key problem as following:1)lack of a unified description model of geospatial data provenance;2)lack of efficient, common mechanism of geospatial data provenance storage and publishing and3)lack of the search of geospatial data provenance based on semantic. In this paper, we provide a geospatial data provenance sharing method based on linked data approach to solve them. The basic idea is:1) to establish the linked geospatial data provenance model for describing the geospatial data provenance and record the semantic information, helping the analysis and integration of geospatial data provenance and supporting for the geospatial data product preservation and interpretation.;2) to achieve the depth sharing and intelligent research of geospatial data provenance based on linked data approach. The main research can be briefly summarized as follow':1) One of the most challenging issues in the geospatial data provenance sharing is lack of uniform describe model to achieve geospatial data provenance sharing and intelligent research in a geospatial service environment. In order to achieve the semantic description and integration of the geospatial data provenance resource, we propose a linked geospatial data provenance ontology integrated with geospatial data type ontology and service type ontology to describe the semantic information, classification feature and hierarchical relationships of geospatial data provenance source.2) We discuss a method for publishing linked geospatial data provenance based on geospatial data provenance service. Firstly, we design the URIs of linked geospatial data provenance and find the URI access method of geospatial data provenance and ontology, Secondly, a mapping model of geospatial data provenance service model and linked geospatial provenance model is proposed and expressed as a mapping file; finally, combination of various linked data publishing method, a publishing system is constructed. We can publish geospatial data provenance and ontology as linked data.3) We study to construct the linkage of linked data sets. We analyze the geospatial data sets in the LOD, and the geospatial data sets are picked up as linked object, if they satisfy the needs of linked geospatial data provenance query and reasoning requirements. The DBpedia is selected as homogeneous linked object of geospatial data provenance and GADM-RDF is selected as relationship linked object. We provide the RDF graph matching algorithm based on Tversky model to construct the homogeneous linkage. The automatic extraction algorithm based on topological relation is proposed to construct the linkage between geospatial data provenance and GADM-RDF. This method can achieve the relationship linkage based on key words matching4) A prototype of linked geospatial data provenance is designed and implemented, which include geospatial data provenance service, linked geospatial data provenance RDF database, middleware for constructing the linking relationship, client and ontology. This prototype verify the above theories and technologiesThe main innovations of this paper can be summarized as below:1) We proposed a geospatial data provenance model to describe the semantic information of geospatial data provenance2) We provided a unified framework and implementation course for publishing linked geospatial data provenance.3) We proposed the RDF graph matching algorithm based on Tversky model to construct the homogeneous linkage and the topological relation extraction algorithm to construct the relationship linkage.