NMF endmember generation method based on abundance distribution constraint

doi:10.7523/j.issn.2095-6134.2020.04.015

Abstract

Abstract: In recent years, the endmember generation method based on non-negative matrix factorization (NMF) attracted much attention. The NMF endmember generation method can be used to obtain endmembers and the abundance matrix simultaneously, and the multiplicative update rule works. Because of the non-convexity of the objective function, NMF endmember extraction easily goes into local extrema. Several constraints were imposed on NMF to alleviate the local extremum problem, but they often broke the multiplicative update rules and increased the processing time. In this work, we propose a new method based on abundance distribution constraint, and the multiplicative iterations can be used. The experimental results show that the method improves the efficiency and accuracy of endmember generation.

Key words: hyperspectral, endmember generation, non-negative matrix factorization (NMF), abundance distribution constraint

CLC Number:

TP751.1

SHI Yue, WANG Hongqi, GUO Xinyi. NMF endmember generation method based on abundance distribution constraint[J]. , 2020, 37(4): 547-552.

References

[1] Boardman J W, Kruse F A, Green R O. Mapping target signatures via partial unmixing of AVIRIS data[J]. Fifth JPL Airborne Earth Science Workshop, 1995,1(1):23-26.
[2] Harsanyi J C, Chang C I. Hyperspectral image classification and dimensionality reduction:an orthogonal subspace projection approach[J]. IEEE Transactions on Geoscience and Remote Sensing, 1994, 32(4):779-785.
[3] Nascimento J M, Dias J M. Vertex component analysis:a fast algorithm to unmix hyperspectral data[J]. IEEE Transactions on Geoscience and Remote Sensing, 2005, 43(4):898-910.
[4] Geng X, Xiao Z, Ji L, et al. A Gaussian elimination based fast endmember extraction algorithm for hyperspectral imagery[J]. ISPRS Journal of Photogrammetry and Remote Sensing, 2013, 79:211-218.
[5] Winter M E. N-FINDR:an algorithm for fast autonomous spectral end-member determination in hyperspectral data[C]//Michael R. Imaging Spectrometry V:Brisbane:International Society for Optics and Photonics, 1999:266-276.
[6] Sun K, Geng X, Wang P, et al. A fast endmember extraction algorithm based on Gram determinant[J]. IEEE Geoscience and Remote Sensing Letters, 2014, 11(6):1124-1128.
[7] Chang C I, Wu C C, Liu W, et al. A new growing method for simplex-based endmember extraction algorithm[J]. IEEE Transactions on Geoscience and Remote Sensing, 2006, 44(10):2804-2819.
[8] Geng X, Ji L, Wang F, et al. Statistical Volume Analysis:a new endmember extraction method for multi/hyperspectral imagery[J]. IEEE Transactions on Geoscience and Remote Sensing, 2016, 54(10):6100-6109.
[9] Craig M D. Minimum-volume transforms for remotely sensed data[J]. IEEE Transactions on Geoscience and Remote Sensing, 1994, 32(3):542-552.
[10] Berman M, Kiiveri H, Lagerstrom R, et al. ICE:a statistical approach to identifying endmembers in hyperspectral images[J]. IEEE Transactions on Geoscience and Remote Sensing, 2004, 42(10):2085-2095.
[11] Zare A, Gader P. Sparsity promoting iterated constrained endmember detection in hyperspectral imagery[J]. IEEE Geoscience and Remote Sensing Letters, 2007, 4(3):446-450.
[12] Geng X, Ji L, Zhao Y, et al. A new endmember generation algorithm based on a geometric optimization model for hyperspectral images[J]. IEEE Geoscience and Remote Sensing Letters, 2013, 10(4):811-815.
[13] Lee D D, Seung H S. Learning the parts of objects by non-negative matrix factorization[J]. Nature, 1999, 401(6755):788.
[14] Cai D, He X, Han J, et al. Graph regularized nonnegative matrix factorization for data representation[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2011, 33(8):1548-1560.
[15] Miao L, Qi H. Endmember extraction from highly mixed data using minimum volume constrained nonnegative matrix factorization[J]. IEEE Transactions on Geoscience and Remote Sensing, 2007, 45(3):765-777.
[16] Jia S, Qian Y. Constrained nonnegative matrix factorization for hyperspectral unmixing[J]. IEEE Transactions on Geoscience and Remote Sensing, 2009, 47(1):161-173.
[17] Wang N, Du B, Zhang L. An endmember dissimilarity constrained non-negative matrix factorization method for hyperspectral unmixing[J]. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2013, 6(2):554-569.
[18] Qian Y, Jia S, Zhou J, et al. Hyperspectral unmixing via L_1/2 sparsity-constrained nonnegative matrix factorization[J]. IEEE Transactions on Geoscience and Remote Sensing, 2011, 49(11):4282-4297.
[19] Wang W, Qian Y. Adaptive L_1/2 sparsity-constrained NMF with half-thresholding algorithm for hyperspectral unmixing[J]. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2015, 8(6):2618-2631.
[20] Cichocki A, Zdunek R, Amari S-i. Hierarchical ALS algorithms for nonnegative matrix and 3D tensor factorization[C]//International Conference on Independent Component Analysis and Signal Separation. Springe, 2007:169-176.
[21] Geng X, Ji L, Yang W, et al. The multiplicative update rule for an extension of the iterative constrained endmembers algorithm[J]. International Journal of Remote Sensing, 2017, 38(23):7457-7467.