A band selection approach based on a modified gray wolf optimizer and weight updating of bands for hyperspectral image

As the curse of dimensionality, band selection is utilized to choose part of feature information from hyperspectral image (HSI) and achieve the highest classification accuracy as possible. It is considered as a non-polynomial hard problem and difficult to be fully solved within the effective period, especially for traditional swarm optimization. Gray wolf optimizer (GWO) is a newly proposed swarm intelligence algorithm based on heuristic search, and chaotic operation helps the algorithm jump out of the local optima. In the paper, a novel band selection approach based on a modified GWO (MGWO) is presented to reduce the data dimension of HSIs, and chaotic operation is utilized to set the index of gray wolves. In addition, the weight of bands is updated by the best and worst fitness values as the iteration of MGWO, and evaluates the quality of each band. Experimental results illustrate that it is better than other state-of-the-art band selection approaches, and satisfied band subsets with the combination from 10% to 25% total number of bands are obtained with higher classification accuracy.