Swarm intelligence: A reliable solution for extracting endmembers from hyperspectral imagery

Abstract

Swarm intelligence has played a more and more important role in solving combinatorial optimization or continuous optimization problems in extracting endmembers from hyperspectral imagery. This paper summarizes the progress of pure pixel-based algorithms using swarm intelligence, and introduces a new minimum volume-based algorithm adopting the artificial bee colony (ABC) model. The experiments with synthetic and real hyperspectral data prove that these swarm intelligence-based algorithms can get comparable or better results than some other state-of-the-art methods. It can be concluded that swarm intelligence is a reliable solution for extracting endmembers from hyperspectral imagery.