Semantic-based approach to spatial data sources integration

  1. MOROCHO ZURITA, CARLOS VILLIE
Supervised by:
  1. Fèlix Saltor Soler Director
  2. Lluís Pérez Vidal Co-director

Defence university: Universitat Politècnica de Catalunya (UPC)

Fecha de defensa: 29 July 2004

Committee:
  1. Mohand-Said Hacid Chair
  2. Albert Abelló Gamazo Secretary
  3. Alfredo Goñi Sarriguren Committee member
  4. José Samos Jiménez Committee member
  5. Jordi Guimet Pereña Committee member

Type: Thesis

Teseo: 104886 DIALNET

Abstract

The integration of information systems, or of data sources, is an urgent need. The quantity of industrial projects and experiences about this subject is evident. While the integration problems on the technical or system level can be considered as solved, the semantic level still needs a substantial amount of research. This thesis aims to tackle this problem area. It is focused on the integration of spatial data sources in order to achieve a solution as solid as possible. Many previous approaches on integration of heterogeneous data sources show that too generic solutions are often impossible to pass from the theory to the practice. Therefore, focusing on a specific application domain this solution may be closer to concrete results. In the first two chapterstwo chapters an overview on the problem of database integration is given and the relevant literature is reviewed. Then on the third chapter three the integration approach, which works on the conceptual level, is introduced, thus the way how to model geographic information is shown and some additional characteristics in a spatial data model are introduced. Chapter four presents theA Variable-Depth and -Level Spatial Ontology (VDLSO), has been designed as part of the Semantic Integration Tool for Spatial Data (SIT-SD) project. In Tthis ontology has been determinesd the main spatial characteristics to be used in the work with spatial objects. A multi level approach on the ontology is considered for facing multi-resolution and multi-representation problems. This is achieved It is getting by means of defined- in- advance views, which could be looked like a pruning process in order to "expand" or "contract" the knowledge in agreement with application needs. Thereforeus, the ontology depth should change from abstract knowledge towards more detailed concepts. Chapter 5 is devoted to the proposal of The a framework proposed on for integration of spatial data sources, a framework whichapproac