Selected Presentations

Home > Presentations > Abstract details

Title

SEMANTIC WEB APPROACH FOR THEMATIC STRUCTURING OF GEOGRAPHIC OBJECTS

Abstract

This presentation aims to show how ontologies can help in better management and diffusion of spatial information. Spatial geometries are graphic representations of concepts just like terms are verbal representations of concepts. Various terms can designate the same concept for a given language (preferred term and synonyms). In the same way, various geometries can represent the same concept according to people, time periods.. Structuring of spatial information through layers doesn't enable to face these heterogeneity. Applications based on Semantic Web languages bring solutions but doesn't cover all these needs (particularly the management of different geometries for a single concept). However it is possible to go further by implementing simple models in opensource tools. Our work reuses existing schemas (to manage semantic information related to concepts, like SKOS) which are expanded by adding new classes or properties in order to manage the variety of terms and geometries needed by users to designate a geographic concept. We will present the generic UML diagrams that have been implemented in our particular application domain : the management of  geographic objects related to marine ecosystems. Our ongoing implementation is entirely based on opensource components (RDF and OWL languages are managed with Jena Semantic Web Framework and a persistent storage in Postgres/Postgis) and demonstrates how it is possible to satisfy different user's point of views (multilinguism, synonyms, various geometries..). Our goal is to reuse this tool to set up different domain ontologies to controll the use of terms or geometries in different information systems (like thesauri into metadata catalogs) accessible through RDF/OWL files or through Web Services. A long term goal would be to agregate the content of such distributed ontologies to set up upper level applications made of local knowledge nodes.

Authors

Julien Barde - IRD (Research for Development Institute), CRHMT
Pascal Cauquil - IRD (Research for Development Institute), CRHMT