Improving Fuzzy Classification by Means of a Segmentation Algorithm

A Spatial Classification Model for Multicriteria Analysis

“…Once the final segmentation based on an n-dimensional aggregation operator has been obtained the result can be processed by an unsupervised classification algorithm like the one designed by Del Amo et al ([8], [10], see also [6], [17], [18]). Let…”

“…In particular, in [6], shows that a pure numerical search, pixel by pixel, may produce useless results, while an approach closer to the analysis of concepts produces clear improvements in the decision process. For example, the search for a concept in [6] was associated to homogeneous regions, and the objective was to obtain a set of basic concepts that would allow a good explanation of the image of study.…”

“…For example, the search for a concept in [6] was associated to homogeneous regions, and the objective was to obtain a set of basic concepts that would allow a good explanation of the image of study. In a first analysis, we got three general classes (natural space, urban area and wetland), a second analysis should be done when a more detail explanation, a higher precision is the ultimate goal for the decision maker.…”