Detection algorithms for hyperspectral imaging applications

TLDR

Hyperspectral imaging faces challenges such as atmospheric effects, spectral variability, mixed pixels, and a need to distinguish classification from detection algorithms. The study develops detection algorithms for full‑pixel and multivariate normal model‑based targets using a likelihood ratio approach. They employ a likelihood ratio framework for full‑pixel detection and use statistical and subspace models to address subpixel detection under spectral variability. Results on real HSI data show algorithm performance and highlight non‑normality of data, suggesting alternative distributions for improved robustness.

Abstract

We introduce key concepts and issues including the effects of atmospheric propagation upon the data, spectral variability, mixed pixels, and the distinction between classification and detection algorithms. Detection algorithms for full pixel targets are developed using the likelihood ratio approach. Subpixel target detection, which is more challenging due to background interference, is pursued using both statistical and subspace models for the description of spectral variability. Finally, we provide some results which illustrate the performance of some detection algorithms using real hyperspectral imaging (HSI) data. Furthermore, we illustrate the potential deviation of HSI data from normality and point to some distributions that may serve in the development of algorithms with better or more robust performance. We therefore focus on detection algorithms that assume multivariate normal distribution models for HSI data.