A variation pixels identification method based on kernel spatial attraction model and local entropy for robust endmember extraction

Abstract

A variation pixels identification method was proposed aiming at depressing the effect of variation pixels, which dilates the theoretical hyperspectral data simplex and misguides volume evaluation of the simplex. With integration of both spatial and spectral information, this method quantitatively defines a variation index for every pixel. The variation index is proportional to pixels local entropy but inversely proportional to pixels kernel spatial attraction. The number of pixels removed was modulated by an artificial threshold factor a. Two real hyperspectral data sets were employed to examine the endmember extraction results. The reconstruction errors of preprocessing data as opposed to the result of original data were compared. The experimental results show that the number of distinct endmembers extracted has increased and the reconstruction error is greatly reduced. 100% is an optional value for the threshold factor a when dealing with no prior knowledge hyperspectral data.

Keywords

variation pixels / hyperspectral / simplex / variation index / local entropy / kernel spatial attraction

References

[1] ChangC IHyperspectral data processing algorithm design and analysis [M], 2013USA, New YorkJohn Wiley and Sons201-206

[2] ChangC I, PlazaA. A fast iterative algorithm for implementation of pixel purity index [J].. Geoscience and Remote Sensing Letters, 2006, 3(1): 63-67

[3] GruningerJ H, RatkowskiA J, HokeM L. The sequential maximum angle convex cone (SMACC) endmember model [C]//. Algorithms for Multispectral, Hyperspectral and Ultra-spectral Imagery, 20043885-3895

[4] HarsanyiJ C, ChangC I. Hyperspectral image classification and dimensionality reduction: An orthogonal subspace projection approach [J].. Geoscience and Remote Sensing, 1994, 32(4): 779-785

[5] NascimentoJ M, Bioucas-DiasJ M. Vertex component analysis: A fast algorithm to unmix hyperspectral data [J].. Geoscience and Remote Sensing, 2005, 43(4): 898-910

[6] WinterM E. N-FINDR: An algorithm for fast autonomous spectral end-member determination in hyperspectral data [C]//. Imaging Spectrometry V, 1999266-275

[7] IordacheM D, Bioucas-DiasJ M, PlazaA. Sparse unmixing of hyperspectral data [J].. Geoscience and Remote Sensing, 2011, 49(6): 2014-2039

[8] MiaoL-d, QiH-rong. Endmember extraction from highly mixed data using minimum volume constrained nonnegative matrix factorization [J].. Geoscience and Remote Sensing, 2007, 45(3): 765-777

[9] MoussaouiS, HauksdottirH, SchmidtF, JuttenC, ChanussotJ, BrieD, DouteS. On the decomposition of Mars hyperspectral data by ICA and Bayesian positive source separation [J]. Neurocomputing, 2008, 71(10): 2194-2208

[10] ChangC I, WuC C, LoC S, ChangM L. Real-time simplex growing algorithms for hyperspectral endmember extraction [J].. Geoscience and Remote Sensing, 2010, 48(4): 1834-1850

[11] HeinzD C, ChangC I. Real-time processing of an unsupervised constrained linear spectral unmixing algorithm [C]//. International Geoscience and Remote Sensing Symposium, 2001372-374

[12] SomersB, AsnerG P, TitsL, CoppinP. Endmember variability in spectral mixture analysis: A review [J].. Remote Sensing of Environment, 2011, 115(7): 1603-1616

[13] KeshavaN, MustardJ F. Spectral unmixing [J].. Signal Processing Magazine, 2002, 19(1): 44-57

[14] ChangC I, WuC C, LiuW M, OuyangY C. A new growing method for simplex-based endmember extraction algorithm [J].. Geoscience and Remote Sensing, 2006, 44(10): 2804-2819

[15] PlazaA, MartinezP, PérezR, PlazaJ. Spatial/spectral endmember extraction by multidimensional morphological operations [J].. Geoscience and Remote Sensing, 2002, 40(9): 2025-2041

[16] LiuJ-m, ZhangJ-she. A new maximum simplex volume method based on householder transformation for endmember extraction [J].. IEEE Transaction on Geoscience and Remote Sensing, 2012, 50(1): 104-118

[17] RezaeiY, MobasheriM R, ZoejM J V, SchaepmanM E. Endmember extraction using a combination of orthogonal projection and genetic algorithm [J].. Geoscience and Remote Sensing Letters, 2012, 9(2): 161-165

[18] ChangC I, WuC C, TsaiC T. Random N-Finder (N-FINDR) endmember extraction algorithms for hyperspectral imagery [J].. Image Processing, 2011, 20(3): 641-656

[19] CallicoG M, LopezS, AguilarB, LopezJ F, SarmientoR. Parallel implementation of the modified vertex component analysis algorithm for hyperspectral unmixing using OpenCL [J].. Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2014, 7(8): 3650-3659

[20] ZhangJ K, RivardB, Sánchez-AzofeifaA, CastroesauK. Intra-and inter-class spectral variability of tropical tree species at La Selva, Costa Rica: Implications for species identification using HYDICE imagery [J].. Remote Sensing of Environment, 2006, 105(2): 129-141

[21] ZareA, HoK C. Endmember variability in hyperspectral analysis: addressing spectral variability during spectral unmixing [J].. Signal Processing Magazine, 2014, 31(1): 95-104

[22] RobertsD A, GardnerM, ChurchR, UstinS, ScheerG, GreenR O. Mapping chaparral in the Santa Monica mountains using multiple endmember spectral mixture models [J].. Remote Sensing of Environment, 1998, 65(3): 267-279

[23] AsnerG, BustamanteM, TownsendA. Scale dependence of biophysical structure in deforested areas bordering the Tapajs national forest [J].. Remote Sensing of Environment, 2003, 87(4): 507-520

[24] BatesonC A, AsnerG P, WessmanC A. Endmember bundles: A new approach to incorporating endmember variability into spectral mixture analysis [J].. Geoscience and Remote Sensing, 2000, 38(2): 1083-1093

[25] JinJ, WangB, ZhangL-ming. A novel approach based on fisher discriminant null space for decomposition of mixed pixels in hyperspectral imagery [J].. Geoscience and Remote Sensing Letters, 2010, 7(4): 699-703

[26] MianjiF A, ZhangY. SVM-based unmixing-to-classification conversion for hyperspectral abundance quantification [J].. Geoscience and Remote Sensing, 2011, 49(11): 4318-4327

[27] CastrodadA, XingZ M, GreerJ B, BoschE, CarinL, SapiroG. Learning discriminative sparse representations for modeling, source separation, and mapping of hyperspectral imagery [J].. Geoscience and Remote Sensing, 2011, 49(11): 4263-4281

[28] MartínG, PlazaA. Spatial-spectral preprocessing prior to endmember identification and unmixing of remotely sensed hyperspectral data [J].. Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2012, 5(2): 380-395

[29] ZorteaM, PlazaA. Spatial preprocessing for endmember extraction [J].. Geoscience and Remote Sensing, 2009, 47(8): 2679-2693

[30] ZorteaM, PlazaA. Spatial-spectral endmember extraction from remotely sensed hyperspectral images using the watershed transformation [C]//. Geoscience and Remote Sensing Symposium, 2010963-966

[31] LopezS, MoureJ F, PlazaA, CallicoG M, LopezJ F, SarmientoR. A new preprocessing technique for fast hyperspectral endmember extraction [J].. Geoscience and Remote Sensing Letters, 2013, 10(5): 1070-1074

[32] MertensK C, BasetsB D, VerbekeL P C, WulfR R D. A sub-pixel mapping algorithm based on sub-pixel/pixel spatial attraction models [J].. International Journal of Remote Sensing, 2006, 27(15): 3293-3310

[33] BezdekJ C. Cluster validity with fuzzy sets [J].. Journal of Cybernetics, 1973, 3(3): 58-73

[34] ChangC I, DuQ. Estimation of number of spectrally distinct signal sources in hyperspectral imagery [J].. Geoscience and Remote Sensing, 2004, 42(3): 608-619