Bayesian Target Detection Algorithms for Solid Subpixel Targets in Hyperspectral Images

We investigate the use of Bayesian methods for hyperspectral subpixel target detection, where uncertainty associated with fill factor is “probabilized” by a suitable prior. Specifically, we present general framework detection employing different models background distribution, comparing choices prior, and investigating numerical schemes evaluating integral. The are furthermore compared to their Generalized Likelihood Ratio Test (GLRT)-based counterparts. Experiments performed over real imagery, both implanted targets, show that incorporating prior knowledge means non-uniform priors emphasizing smaller factors outperforms usage “noninformative” uniform enhances Bayes performance beyond GLRT, result observed parametric non-parametric models. find even “rough” can successfully leverage context-based information sizes most interest. further observe Gauss-Legendre integration scheme provides efficient integral approximation while maintaining desirable admissibility property methods.