The aim of this project is to design a new module for the firm's raster-based Image Processing and Geographic Information System (GIS) software. The concept will be used to enhance standard digital vector land-use maps using a knowledge-based expert system that accesses available GIS data sources, such as satellite imaging, topographic data, soil data, and hydrology. The focus will be on an interface to an existing expert-system shell. The intention is to build on previous studies with rule-based land-use classification. The knowledge-based module will query the user about the rules decision, store and modify the decision rules, and write the macro needed to decide the final land-use classification. The innovation of the program will be to enhance today's image-processing techniques in cases where the land-use classification will be determined by a polygon-specific, and not pixel-specific, classifier. The module will rely on spatial rasters and auxiliary data to duplicate products now produced by traditional airphoto interpretation. The classification rules will be developed using the firm's raster-based GIS.Classified satellite imagery may be put into a GIS database fast and with less reliance on scientific understanding of spectral analysis. This innovation will significantly reduce the time and effort spent putting land-use data into a digital database and increase the analytic capability of GIS technology.STATUS: Phase I Only
