Hyperspectral band selection using firefly algorithm

A novel band selection algorithm for hyperspectral dimensionality reduction by improving the firefly algorithm is put forward. Specifically, the framework which using bio-inspired algorithm for hyperspectral band selection is described; the between-class separability criteria such as Jeffreys-Matusita (JM) distance, transformation divergence (TD) are used for objective function. A HYDICE Washington DC Mall data was used in the experiments, and the results have proved the promising ability of the proposed method for hyperspectral band selection.