A fuzzy classification of sub-urban land cover from remotely sensed imagery

Abstract

Fuzzy methods in remote sensing have received growing interest for their particular value in situations where the geographical phenomena are inherently fuzzy. A fuzzy approach is investigated for the classification of sub-urban land cover from remote sensing imagery and the evaluation of classification accuracy. Under the fuzzy strategy, fuzziness, intrinsic to both remotely sensed data and ground data, is accommodated and usefully explored. For comparative purposes, hard and fuzzy classifications were produced and tested using hard and fuzzy evaluation techniques. The results show that the fuzzy approach holds advantages over both conventional hard methods and partially fuzzy approaches, in which fuzziness in only the remotely sensed imagery is accommodated. It was found that Kappa coefficients were more than doubled when applying the fuzzy evaluation technique as opposed to the hard evaluation technique. Furthermore, the fuzzy approach paves the way towards an integrated handling of remotely sensed data and other spatial data.