ONTOLOGY-BASED MULTI-SOURCE DATA INTEGRATIONFOR DIGITAL SOIL MAPPING

摘要

There is a need for cheap methods for digital soil mapping on intermediate scales that make optimal use of existing multi-source datasets on both general and detailed scales. Apart from the spatial challenges that often have to be faced in map integration the semantics of datasets have to be well understood for successful data integration. So-called "ontology-based" approaches for the semantic integration of multi-source geographical datasets may be used to give a firm conceptual basis to digital soil mapping from multi-scale, multi-source geographic data. This chapter explores the use of ontologies in se-mantic data integration. A first version of an approach for ontology-based data integra-tion for soil-landscape mapping is presented consisting of semantic factoring, ontology definition, reference model construction and data integration. This approach is illustrated with the semantic integration of a small-scale (1:400,000) soil-geomorphic map with a geological map at 1:50,000 scale of the Antequera area in Spain. Comparison of the results of our approach to semantic integration with an Antequera geo-pedological legend de-signed to map soil-landforms in this area at 1:50,000 scale shows a clear correspondence of ontologies.