Hyperspectral band selection based on the combination of multiple classifiers

Due to the high dimensionality of hyperspectral data, dimension reduction is becoming an important problem in hyperspectral image classification. Band selection can retain the information which is capable of keeping the original meaning of the data, and thus has attracted more attention. This paper tackles the band selection problem from the perspective of multiple classifiers combination, which can obtain higher classification accuracy. The proposed approach is made up of three steps. First, support vector machine (SVM) and genetic algorithm are employed to search for several groups of initial band subset, on which a pool of classifiers is constructed. Second, a new classifier selection algorithm based on error diversity is used to select several member classifiers from the initial classifier pool. And finally the classification is performed through dynamic classifier selection based on local classification accuracy. The experimental results on the Indian Pine benchmark data set show that the new method can select those bands with more discriminative information and improve the classification performance effectively.