Binary Coronavirus Disease Optimization Algorithm for Spectral Band Selection

Abstract

Remote sensing has become a very interesting tool in various applications because its ability to capture detailed spectral information with a wide range of wavelengths. However, the large dimensionality of hyperspectral image and the inherent complexity provides many challenges for image classification and the accuracy rate. One of the critical preprocessing step in wealth of data hyperspectral image classification is the spectral band selection which play a very important role to reduce the dimensionality and by consequences reduce the Huge phenomena. In this paper, we propose a new spectral band selection approach based on a new metaheuristic called Coronavirus Disease Optimization Algorithm (COVIDOA). A binary version of this metaheuristic is proposed with a new objective function based on accuracy rate and distance measure. The proposed approach will be tested on three hyperspectral images widely used in the literature and demonstrating its efficacy in improving classification accuracy rate.