Band Selection Method Based on Orthogonal Projection and Cross Entropy

Aiming at the problem of insufficient information or excessive redundancy of bands selected by a large number of band selection methods at present, this paper studies the band selection method based on the fusion of information and redundancy on the basis of the maximum empirical volume (MEV) theory. A band selection method based on orthogonal projection and cross entropy is proposed by combining the two indexes to evaluate the bands. The theoretical basis of the method is described and the process of the proposed method is analyzed in depth. Finally, the advantages of the proposed method are verified through experiments and comparisons on three datasets. The proposed method has the advantages of large amount of information in the selected bands, low redundancy between bands, high computational efficiency and insensitivity to noise bands.