Hyperspectral Data Processing: Algorithm Design and Analysis

27

Variable-Number Variable-Band Selection for Hyperspectral Signals

This chapter presents a novel band selection-based feature characterization technique for a hyperspectral signature vector, referred to as variable-number variable-band selection (VNVBS). Since a hyperspectral signature vector is generally characterized by its spectral profile, its feature characterization can be achieved by selecting appropriate bands from the original set of spectral bands, and the number of bands to be selected is totally determined by its original spectral profile. As a result, two hyperspectral signature vectors may require different sets of bands for spectral feature characterization. Therefore, VNVBS allows users to select a different number of variable bands in accordance with spectral characteristics of a hyperspectral signature vector. In order for VNVBS to select an appropriate subset of bands for a hyperspectral signature vector, a new band prioritization criterion, referred to as orthogonal subspace projector-based band prioritization criterion (OSP-BPC), is derived. It assigns a different priority score to each spectral band of a hyperspectral signature vector such that various features can be captured by VNVBS. Accordingly, VNVBS can be interpreted as a spectral band selection-based feature extraction technique for hyperspectral signature characterization.