A New Band Selection Strategy for Target Detection in Hyperspectral Images

This manuscript presents a novel methodology for band selection (BS) for hyperspectral sensors (HSs) tailored to target detection applications. The new selection strategy chooses a subset of bands that maximizes an objective function suitable for target detection. In particular, it extracts the subset of bands that optimizes the probability of detecting a target (P D ) in a given background, for a fixed probability of false alarm (P FA ). An experimental example of the methodology effectiveness is given. In the example the well-known Adaptive Matched Filter (AMF) detector and synthetic data derived from an AVIRIS hyperspectral image are considered. The results obtained show that the the new strategy outperforms two existing BS algorithms.