A Nature-inspired Framework for Hyperspectral Band Selection

Although hyperspectral images acquired by onboard satellites provide information from a wide range of wavelengths in the spectrum, the obtained information is usually highly correlated. This article proposes a novel framework to reduce the computation cost for large amount of data based on the efficiency of the Optimum-Path Forest classifier and the power of meta-heuristic algorithms to solve combinatorial optimizations. Simulations on two public datasets have shown that the proposed framework can indeed improve the effectiveness of the Optimum-Path Forest and considerably reduce data storage costs.