Mapping land cover types as fuzzy sets

Researches have suggested that fuzzy sets form a more appropriate basis for land cover mapping than traditional Boolean classification. However, to give a crisp answer of which land cover types pixels in remote sensing images belong to, fuzzy clustering methods such as FMC always lost some subtle vague information. In this paper, the result of the Boolean classification using a fuzzy clustering method was analyzed and based on the result of the classification, we tried to map land cover types as type 1 fuzzy sets and type 2 fuzzy sets. Results show that type 1 fuzzy sets and type 2 fuzzy sets kept more subtle information that was lost in the processing of converting the fuzzy classification result to a Boolean one. And the structure of the study area can be more explicit represented.