Title

A distributed mobile agent conflation model utilizing an image change detection algorithm

Date of Award

2002

Degree Type

Dissertation

Degree Name

Doctor of Philosophy (PhD)

Department

Computing

First Advisor

Maria Cobb

Advisor Department

Computing

Abstract

The Geographic Information System (GIS) is an integrated technology that incorporates concepts from computer graphics, spatial modeling, and database management. The distributed intelligent mobile agent technique, which successfully incorporated more powerful technology, is becoming an important issue in Geographical Information Systems. Within the distributed environment, the compatibilities and consistencies are mainly concerned issues. The "conflation" is an important and challenging technique to handle these issues. Generally, conflation means to combine information from different sources and then to produce better information. Up to now, the conflation consideration has become much broader in GIS research fields. Many efforts have been made. However, conflation is still a challenging research field due to the complexity of real applications. Seen from the existing conflation paradigms, the conflation algorithms have been ad hoc, designed for specific purposes. The focus of the dissertation is placed mainly on processing the vector-based conflation problems. Considered the vulnerability to deal with time in existing conflation algorithms, the endeavor of the dissertation is to explore the ways in which conflation capabilities can be augmented with the aid of change detection techniques. A general and flexible conflation model is proposed for the distributed mobile agent systems. Based on the model, over time considerable effort is specially spent on the development of image change detection algorithm. Since image change detection, like many other applications in GIS, requires to handle fuzziness and uncertainty, an innovation is investigated--it is an intelligent approach in which the issue associated with fuzziness and uncertainty has been tackled by introducing a Certainty Factor. A hierarchical structure for the fuzzy inference is figured out. Theoretical analysis and real image evaluation show that it can provide significant results.