Abstract
Hyperspectral imaging can be used in assessing the quality of foods by decomposing the image into constituents such as protein, starch, and water. Observed data can be considered a mixture of underlying characteristic spectra (endmembers), and estimating the constituents and their abundances requires efficient algorithms for spectral unmixing. We present a Bayesian spectral unmixing algorithm employing a volume constraint and propose an inference procedure based on Gibbs sampling. We evaluate the method on synthetic and real hyperspectral data of wheat kernels. Results show that our method perform as good or better than existing volume constrained methods. Further, our method gives credible intervals for the endmembers and abundances, which allows us to asses the confidence of the results.
Similar content being viewed by others
Notes
In the literature, the constraints that abundances must sum to one is sometimes refered to as an additivity constraint.
In the notation used in this paper, matrices and vectors are denoted by capital and lower case bold letters respectively. Two subscripts denotes a sub matrix or sub vector with the corresponding rows and columns, where a colon denotes all indices, and \(\tilde m\) denotes all indices except m. For example, w :k denotes the kth column of W and \(\mathbf{w}_{m\tilde k}\) denotes the mth row of W with the kth element removed. A single element of the matrix W is denoted by w mk .
For simplicity, we use a slightly different notation in this section.
In practice, of course, we cannot generate a random number from this improper prior, so we can think of generating w from a uniform distribution between zero and u w , where u w is some very large number.
All scatterplots in this paper are presented as subspace projections onto the first and second principal component (PC) based on the data points.
References
Arngren, M., Schmidt, M., & Larsen, J. (2009). Bayesian nonnegative matrix factorization with volume prior for unmixing of hyperspectral images. In Machine learning for signal processing, IEEE workshop on (MLSP). doi:10.1109/MLSP.2009.5306262.
Arngren, M., Schmidt, M. N., & Larsen, J. (2010). Unmixing of hyperspectral images using Bayesian nonnegative matrix factorization with volume prior. Technical note. http://www.imm.dtu.dk/pubdb/p.php?5836.
Arngren, M., Schmidt, M. N., & Larsen, J. (2010). Unmixing of hyperspectral images using bayesian nonnegative matrix factorization with volume prior, Matlab toolbox. http://www.imm.dtu.dk/pubdb/p.php?5834.
Berman, M., Kiiveri, H., Lagerstrom, R., Ernst, A., Dunne, R., & Huntington, J. (2004). ICE: A statistical approach to identifying endmembers in hyperspectral images. IEEE Transactions on Geoscience and Remote Sensing, 42(10), 2085–2095.
Bioucas-Dias, J., & Nascimento, J. (2008). Hyperspectral subspace identification. IEEE Transactions on Geoscience and Remote Sensing, 46(8), 2435–2445.
Boardman, J. W., Kruse, F. A., & Green, R. O. (1995). Mapping target signatures via partial unmixing of aviris data. In Summaries of JPL airborne earth science workshop.
Chan, T. H., Chi, C. Y., Huang, Y. M., & Ma, W. K. (2009). A convex analysis-based minimum-volume enclosing simplex algorithm for hyperspectral unmixing. IEEE Transactions on Signal Processing, 57(11), 4418–4432.
Dobigeon, N., Moussaoui, S., Coulon, M., Tourneret, J. Y., & Hero, A. O. (2009). Joint Bayesian endmember extraction and linear unmixing for hyperspectral imagery. IEEE Transactions on Signal Processings, 57(11), 4355–4368.
Dobigeon, N., Moussaoui, S., Tourneret, J. Y. Y., & Carteret, C. (2009). Bayesian separation of spectral sources under non-negativity and full additivity constraints. Signal Processing, 89(12), 2657–2669.
Dobigeon, N., & Tourneret, J. Y. (2007). Efficient sampling according to a multivariate Gaussian distribution truncated on a simplex. Tech. Rep., IRIT/ENSEEIHT/TeSA.
Eches, O., Dobigeon, N., & Tourneret, J. Y. (2010). Estimating the number of endmembers in hyperspectral images using the normal compositional model and a hierarchical Bayesian algorithm. IEEE Journal of Selected Topics in Signal Processing, 4(3), 582–591.
Gelfand, A. E., & Smith, A. F. M. (1990). Sampling-based approaches to calculating marginal densities. Journal of the American Statistical Association, 85(410), 398–401.
Geman, S., & Geman, D. (1984). Stochastic relaxation, gibbs distributions, and the Bayesian restoration of images. IEEE Transactions on Pattern Analysis and Machine Intelligence, 6, 721–741. doi:10.1109/TPAMI.1984.4767596.
Geweke, J. (1991). Efficient simulation from the multivariate normal and student-t distributions subject to linear constraints and the evaluation of constraint probabilities. In Proceedings the 23rd symposium on the interface between computer sciences and statistics (pp. 571–578). doi:10.1.1.26.6892.
Hoyer, P. O. (2002). Non-negative sparse coding. Neural Networks for Signal Processing, 7, 557–565.
Masalmah, Y. M. (2007). Unsupervised unmixing of hyperspectral imagery using the constrained positive matrix factorization. Ph.D. thesis, Computing And Information Science And Engineering, University Of Puerto Rico, Mayagez Campus.
Mazet, V., Brie, D., & Idier, J. (2005). Simulation of postive normal variables using several proposal distributions. In IEEE workshop on statistical signal processing (pp. 37–42). doi:10.1109/SSP.2005.1628560.
Miao, L., & Qi, H. (2007). Endmember extraction from highly mixed data using minimum volume constrained nonnegative matrix factorization. IEEE Transactions on Geoscience and Remote Sensing, 45(3), 765–777.
Moussaoui, S., Brie, D., Mohammad-Djafari, A., & Carteret, C. (2006). Separation of non-negative mixture of non-negative sources using a Bayesian approach and mcmc sampling. IEEE Transactions on Signal Processing, 54(11), 4133–4145.
Nascimento, J. M. P., & Dias, J. M. B. (2005). Vertex component analysis: A fast algorithm to unmix hyperspectral data. IEEE Transactions on Geoscience and Remote Sensing, 43(4), 898–910.
Ochs, M. F., Stoyanova, R. S., Arias-Mendoza, F., & Brown, T. R. (1999). A new method for spectral decomposition using a bilinear bayesian approach. Journal of Magnetic Resonance, 137(1), 161–176.
Paatero, P., & Tapper, U. (1994). Positive matrix factorization: A non-negative factor model with optimal utilization of error estimates of data values. Environmetricsm 5(2), 111–126.
Parra, L., Spence, C., Sajda, P., Ziehe, A., & Mller, K. R. (1999). Unmixing hyperspectral data. Neural Information Processing Systems, 12, 942–948.
Pauca, V. P., Piper, J., & Plemmons, R. J. (2006). Nonnegative matrix factorization for spectral data analysis. Linear Algebra and Its Applications, 416(1), 29–47.
Plaza, A., Martnez, P., Prez, R., & Plaza, J. (2004). A quantitative and comparative analysis of endmember extraction algorithms from hyperspectral data. IEEE Transactions on Geoscience and Remote Sensing, 42(3), 650–663.
Sajda, P., Du, S., & Parra, L. (2003). Recovery of constituent spectra using non-negative matrix factorization. Proceedings of the SPIE - The International Society for Optical Engineering, 5207(1), 321–331.
Schachtner, R., Pppel, G., Tom, A. M., & Lang, E. W. (2009). Minimum determinant constraint for non-negative matrix factorization. Lecture Notes in Computer Science, 5441/2009, 106–113.
Schmidt, M. N. (2009). Linearly constrained bayesian matrix factorization for blind source separation. In Y. Bengio, D. Schuurmans, J. Lafferty, C. K. I. Williams, & A. Culotta (Eds.), Advances in neural information processing systems (Vol. 22, pp. 1624–1632).
Schmidt, M. N., & Laurberg, H. (2008). Nonnegative matrix factorization with Gaussian process priors. Computational Intelligence and Neuroscience p. Article ID 361705.
Schmidt, M. N., Winther, O., & Hansen, L. K. (2009). Bayesian non-negative matrix factorization. In Independent component analysis and signal separation, international conference on lecture notes in computer science (LNCS) (Vol. 5441, pp. 540–547).
Winter, M. E. (1999). N-findr: An algorithm for fast autonomous spectral end-member determination in hyperspectral data. Proceedings of SPIE - The International Society for Optical Engineering, 3753, 266–275.
Author information
Authors and Affiliations
Corresponding author
Rights and permissions
About this article
Cite this article
Arngren, M., Schmidt, M.N. & Larsen, J. Unmixing of Hyperspectral Images using Bayesian Non-negative Matrix Factorization with Volume Prior. J Sign Process Syst 65, 479–496 (2011). https://doi.org/10.1007/s11265-010-0533-2
Received:
Revised:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s11265-010-0533-2