Search

A hierarchical Bayesian model is presented for hyperspectral image unmixing. An extension of the algorithm is studied for mixtures with unknown numbers of spectral components belonging to a known library.

A joint segmentation algorithm is proposed for piecewise constant autoregressive (AR) processes recorded by several independent sensors. The results are illustrated by several simulations conducted with synthetic signals and real arc-tracking and speech signals.

Bayesian sampling algorithms are studied for segmenting single and multiple time series by obeying Poisson distributions with piecewise constant parameters. A Gibbs sampling strategy has allowed joint estimation of the unknown parameters and hyperparameters and the results obtained from synthetic and real photon counting data have shown the performance of the proposed algorithm.

This paper studies a fully Bayesian algorithm for endmember extraction and abundance estimation for hyperspectral imagery. Each pixel of the hyperspectral image is decomposed as a linear combination of pure endmember spectra following the linear mixing model. The estimation of the unknown endmember spectra is conducted in a unified manner by generating the posterior distribution of abundances and endmember parameters under a hierarchical Bayesian model. This model assumes conjugate prior distributions for these parameters, accounts for nonnegativity and fulladditivity constraints, and exploits the fact that the endmember proportions lie on a lower dimensional simplex. A Gibbs sampler is proposed to overcome the complexity of evaluating the resulting posterior distribution.. This sampler generates samples distributed according to the posterior distribution and estimates the unknown parameters using these generated samples. The accuracy of the joint Bayesian estimator is illustrated by simulations conducted on synthetic and real AVIRIS images. [ABSTRACT FROM AUTHOR]