Data and Quality Integration in GIS
Résumé
Decision making often relies on digital geographic data. These decisions are sometimes very important, and their consequences may be very costly. On one hand, any decision is affected by the quality of the data, and, in spite of the information about data quality, which is provided in metadata by the Mapping Agencies, almost nobody uses it. The reason is identified - a viewpoint mismatch - but no solution has been elaborated to date. On the other hand, there is a need to integrate several data sources with multiple definitions organized in different schemas, which match imperfectly and whose data quality can be very heterogeneous. The work described in this paper aims to provide a data integration approach that copes with quality information. The problem consists on integrating different heterogeneous and autonomous geographic data sources having different qualities. Our goal is to define a mediation approach in order to integrate not only the data sources but also the tools (programs) that manipulate them as well as their qualities. The originality of this work consists on a new way of resolving the problem of data integration with taking account of the quality information and tools, notably quality processing ones.