MCMC algorithm for spectral unmixing
1 : Institut de recherche en informatique de Toulouse
(IRIT)
Université Toulouse le Mirail - Toulouse II, Université des Sciences Sociales - Toulouse I, CNRS : UMR5505, Université Paul Sabatier - Toulouse III
Spectral unmixing is a crucial step while analyzing hyperspectral images in remote sensing and astronomical applications. It consists of decomposing measured pixels into a set of elementary spectra and quantifying their respective proportions in the observed mixtures. Unsupervised spectral unmixing can be formulated as a blind source separation problem. Bayesian estimation offers a convenient framework to tackle this problem. This talk presents a Bayesian estimation algorithm derived to recover the parameters of interest (spectral components and corresponding mixing coefficients). It is based on the wide assumption of a linear observation model. The physical constraints implied by this model are naturally ensured within the Bayesian framework by assigning appropriate priors to the unknown parameters. Inference is conducted using a Markov chain Monte Carlo algorithm. Examples on real hyperspectral data will be proposed.
Keywords: spectral unmixing, hyperspectral images, Bayesian inference, MCMC algorithm.
References:
[1] N. Dobigeon, S. Moussaoui, M. Coulon, J.-Y. Tourneret and A. O. Hero, "Joint Bayesian endmember extraction and linear unmixing for hyperspectral imagery," IEEE Trans. Signal Processing, vol. 57, no. 11, pp. 4355-4368, Nov. 2009.
[2] N. Dobigeon, S. Moussaoui, J.-Y. Tourneret and C. Carteret, "Bayesian separation of spectral sources under non-negativity and full additivity constraints", Signal Processing, vol. 89, no. 12, pp. 2657-2669, Dec. 2009.
[3] F. Schmidt, M. Guiheneuf, S. Moussaoui, E Tréguier, A. Schmidt and N. Dobigeon, "Implementation strategies for hyperspectral unmixing using Bayesian source separation," IEEE Trans. Geoscience and Remote Sensing, vol. 48, no. 11, pp. 4003-4013, Nov. 2010.
Keywords: spectral unmixing, hyperspectral images, Bayesian inference, MCMC algorithm.
References:
[1] N. Dobigeon, S. Moussaoui, M. Coulon, J.-Y. Tourneret and A. O. Hero, "Joint Bayesian endmember extraction and linear unmixing for hyperspectral imagery," IEEE Trans. Signal Processing, vol. 57, no. 11, pp. 4355-4368, Nov. 2009.
[2] N. Dobigeon, S. Moussaoui, J.-Y. Tourneret and C. Carteret, "Bayesian separation of spectral sources under non-negativity and full additivity constraints", Signal Processing, vol. 89, no. 12, pp. 2657-2669, Dec. 2009.
[3] F. Schmidt, M. Guiheneuf, S. Moussaoui, E Tréguier, A. Schmidt and N. Dobigeon, "Implementation strategies for hyperspectral unmixing using Bayesian source separation," IEEE Trans. Geoscience and Remote Sensing, vol. 48, no. 11, pp. 4003-4013, Nov. 2010.
