Geographic data are important sources of information in commerce, industry, and everyday life. There is a growing awareness -will grow even more with the future progress of interoperability- that geographic information is often imprecise, uncertain, incomplete, qualitative in nature and difficult to combine with quantitative information in current GIS. Hence, we aim to provide theories and tools to help GI users " to be rigorous with uncertain information". This project continues the Esprit-4 project REVIGIS, which has shown the feasibility of using a fusion of fuzzy logics, non-monotonic logics and rough sets in such situations. Phase2 will develop into a collection of industry-strength techniques, demonstrated in large-scale applications with a mix of partners in academia, public agencies and commerce. It will also develop these techniques for application in the case where the knowledge base is derived from distributed and semantically heterogeneous spatial databases.
The previous REVIGIS (Esprit-4) project explored the potential of the use of the fusion of different formal approaches to handling spatial data uncertainty.
The objectives are now:
- To further develop techniques for the smooth integration of new information (data or constraints) into geographic databases;
- To explicitly represent and reason with spatial uncertainty in systems;
- To integrate qualitative techniques (developed in Phase 1) for dynamic spatial uncertainty handling and representation into state-of-the-art systems;
- To evaluate algorithmic complexity and develop efficient algorithms for reasoning with spatial uncertainty;
- To develop these techniques in the context of environmental and cadastral applications.
DESCRIPTION OF WORK
This work continues the Esprit4-LTR project REVIGIS (Phase 1) by further developing the techniques for representing and reasoning about " knowledge revision " in spatial datasets, and integrating the techniques into current commercial and scientific applications. Phase 1 showed the feasibility of providing the foundations for reasoning in such situations, and showed the useful roles to be played of fuzzy logic, non-monotonic logics and rough sets. Phase 1 provided a unified approach to such methods, and developed criteria for the assessment of which technique to apply to which application. Phase 2 will develop the work from a demonstrator of the success of the foundation to a collection of industry-strength techniques, demonstrated in large-scale applications with a mix of partners in academia, commerce and national scientific organizations. The project will also develop these techniques for application in the case (now usual) where the knowledge base is derived from distributed and semantically heterogeneous spatial databases, including the information extraction from primary data, such as remotely sensed imagery. This work is enhanced by the addition of new commercial and academic partners. Consequently, this requires evaluation of the impact of this work on national and European data quality standards. The project will prove the applicability of concepts, by constructing tools for application to the environmental and cadastral information bases. The theoretical and long-term foundational research strand will also be further developed.
The workplan divides the project into packages:
- Reasoning under uncertainty with geographic knowledge;
- Land description with ill-defined spatial objects;
- Fusion and revision of strongly constrained spatial objects;
- Scientific general coordination, project management and information dissemination: http://www.cmi.univ-mrs.fr/REVIGIS/
Funding SchemeCSC - Cost-sharing contracts
ST5 5BG Staffordshire
7514 AE Enschede
83957 La Garde Cedex
G1K 7P4 Sainte-foy, Quebec