Land cover classification with an expert system method using landsat ETM data

Title of thesis : Land Cover Classification with an Expert System Method Using LandsatETM DataAuthor : Oğuzhan KAHYAABSTRACTExpert systems allow for the integration of remote sensed data with other sources ofgeoreferenced information (such as land use data, spatial texture, and digital elevation model(DEMs) to obtain greater classification accuracy. The spatial and temporal distribution of landcover is a fundamental dataset for urban ecological research.In this study, the design and implementation of an expert system for image classification isdiscussed. Logical decision rules are used with the various datasets to assign class values toeach pixel.Expert system is very suitable for the work of image interpretation as a powerful means ofinformation integration. ETM data acquired in 2000 were initially classified into seven classesfor land cover using a maximum likelihood decision rule. In addition, spatial texture of theETM data was calculated. An expert system was constructed to perform postclassificationsorting of the initial land cover classification using additional spatial datasets (such as texture,land use data).The overall accuracy of expert classification was 95.80%. Individual class accuracy rangedfrom 75% to 100% for each class. The methodology described here will be used to monitorand determine future land cover changes.Key Words: Image classification, Expert system, Image interpretation, Expert classification