Supervised fusion-classification of multispectral images using fuzzy sets theory

Abstract

A new methodology is proposed for supervised fusion-classification of multispectral images based on fuzzy set theory. The method is suited to mapping land-cover in a highly complex landscape. As the fuzzy set theory is intrinsically suited for dealing with the mixed pixels problem and is able to represent ill-defined classes in a natural way, the proposed method overcomes the drawbacks of conventional statistical classification methods. The uncertainty associated with multispectral data is reduced while the imprecise information of the multispectral image is explicitly measured and integrated in the proposed fusion-classification decision rule. The effectiveness of the decision rule in reducing the rate of miss-classification is then proved. We apply our methodology for the problem of classifying two different complex scenes: Laghouat City and its periphery in S Algeria, using a multispectral image provided by Landsat-TM, and Djebel-Amour and its periphery in SW Algeria, using a multispectral image provided by SPOT.