Fuzzy approaches for improving a classification accuracy in remote sensing image data

The study proposes a fuzzy analytic algorithm for classification of mixed pixel (mixel) image. First, the misclassified problem and information loss of mixed pixels are examined. To solve the mixel's uncertainty, a fuzzy function which produces a fuzzy number is suggested. Second, algorithms for the analytic fuzzy approach are represented. This analytic method is applied to two bridges across the Hon River and an area adjacent to the river in Seoul to verify it. >