An Approach for Data Fusion Using Uncertain Knowledge in Geographical Information Systems

Abstract

The digital terrain maps become widespread, and Geographical Information Systems (GIS) have come into the limelight. One of the key technologies needed in GIS is data fusion reasoning. The function of data fusion is to consider various geographical data, such as "road" and "slope", and to make total evaluation. In this paper, we propose a data fusion reasoning technology using uncertain knowledge. Data fusion knowledge contains some uncertainty. For example, the knowledge for evaluating mobility costs is uncertain because it is qualitative, such as "we want to refuse the steep place.". We introduced two uncertainty reasoning mecha nisms to represent such data fusion process. One is fuzzy reasoning, and the other is Dempster-Shafer theory. We also offer knowledge editing facilities for describing data fusion knowledge, such as a data-flow diagram editor for designing data fusion process and a membership function editor for describing data abstraction methods. These knowledge editors facilitate the development and modification of data fusion knowledge base for GIS.