Nonlinear Unmixing of Hyperspectral Images: Models and Algorithms
Open accessThe DOI matches its Crossref record; 55 of the 57 references it cites were found in a registry or an index.
Abstract
When considering the problem of unmixing hyperspectral images, most of the literature in the geoscience and image processing areas relies on the widely used linear mixing model (LMM). However, the LMM may be not valid, and other nonlinear models need to be considered, for instance, when there are multiscattering effects or intimate interactions. Consequently, over the last few years, several significant contributions have been proposed to overcome the limitations inherent in the LMM. In this article, we present an overview of recent advances in nonlinear unmixing modeling.
Overview
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
The cited source says what the claim attributes to it.
In the citing paper
“However, when nonlinear interactions become significant, more intricate models are required (96).”
From Hyperspectral unmixing for Raman spectroscopy via physics-constrained autoencoders (2024) · cites this paper as [96]
The cited source says what the claim attributes to it.
In the citing paper
“However, accounting for nonlinear mixing interactions increases the complexity and computational cost of unmixing (96, 97), an issue of especial relevance in Raman spectroscopy where datasets (e.g., imaging/volumetric scans) are typically larger than in remote sensing.”
From Hyperspectral unmixing for Raman spectroscopy via physics-constrained autoencoders (2024) · cites this paper as [96]
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
Citations
Cites
Hyperspectral Unmixing Overview: Geometrical, Statistical, and Sparse Regression-Based Approaches
Bioucas-Dias, J. M., Plaza, A., Dobigeon, N., Parente, M., Du, Q., Gader, P., & Chanussot, J. (2012) · IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing · doi:10.1109/jstars.2012.2194696 · Cited by 2,836 · cited 4×
“In [1], it is explained that linear mixtures are reasonable when two assumptions are wholly fulfilled.”
BackgroundMethodInfluentialKeshava, N., & Mustard, J. F. (2002) · IEEE Signal Processing Magazine · doi:10.1109/79.974727 · Cited by 2,549 · cited 2×
“More than 10 years after Keshava and Mustard’s comprehensive review article on spectral unmixing [11], this article provides an updated review focusing on non-linear unmixing techniques introduced in the past decade.”
BackgroundReflectance spectroscopy: Quantitative analysis techniques for remote sensing applications
Clark, R. N., & Roush, T. L. (1984) · Journal of Geophysical Research: Solid Earth · doi:10.1029/jb089ib07p06329 · Cited by 2,055 · cited 1×
“Secondly, the photons that reach the sensor must interact with only one material, as is the case in checkerboard type scenes [13].”
BackgroundBidirectional reflectance spectroscopy: 1. Theory
Hapke, B. (1981) · Journal of Geophysical Research: Solid Earth · doi:10.1029/jb086ib04p03039 · Cited by 1,888 · cited 3×
“In [2], the author derives an analytical model to express the measured reflectances as a function of parameters intrinsic to the mixtures, e.”
BackgroundMethodInfluentialTheory of Reflectance and Emittance Spectroscopy
Hapke, B. (1993) · Cambridge University Press eBooks · doi:10.1017/cbo9780511524998 · Cited by 1,849 · cited 4×
“The abundances can be interpreted as the relative areas occupied by the materials in a given image pixel [3].”
BackgroundInfluentialHeinz, D. C., & Chein-I-Chang (2001) · IEEE Transactions on Geoscience and Remote Sensing · doi:10.1109/36.911111 · Cited by 1,847 · cited 2×
“The fully constrained least square (FCLS) algorithm [35] was used for linear mixtures.”
MethodThe discrete-dipole approximation and its application to interstellar graphite grains
Draine, B. T. (1988) · The Astrophysical Journal · doi:10.1086/166795 · Cited by 1,528 · cited 1×
“Other popular approximating models include the discrete-dipole approximation [16] and the Shkuratov’s model [17] (interested readers are invited”
MethodA Model of Spectral Albedo of Particulate Surfaces: Implications for Optical Properties of the Moon
Shkuratov, Y., Starukhina, L., Hoffmann, H., & Arnold, G. (1999) · Icarus · doi:10.1006/icar.1998.6035 · Cited by 471
A Signal Processing Perspective on Hyperspectral Unmixing: Insights from Remote Sensing
Ma, W.-K., Bioucas-Dias, J. M., Chan, T.-H., Gillis, N., Gader, P., Plaza, A. J., Ambikapathi, A., & Chi, C.-Y. (2014) · IEEE Signal Processing Magazine · doi:10.1109/msp.2013.2279731 · Cited by 456
Nonlinear Unmixing of Hyperspectral Images Using a Generalized Bilinear Model
Halimi, A., Altmann, Y., Dobigeon, N., & Tourneret, J.-Y. (2011) · IEEE Transactions on Geoscience and Remote Sensing · doi:10.1109/tgrs.2010.2098414 · Cited by 443 · cited 2×
“linear model (GBM) has been proposed in [22] by setting βi,j,p = γi,j,pai,paj,p”
BackgroundMethodSelf-citation
Research works printed but not linked: 6
- [30] J. Plaza, P. Martinez, R. Perez, and A. Plaza, 'Nonlinear neural network mixture models for fractional abundance estimation in AVIRIS hyperspectral images,' in Proc. XIII NASA/Jet Propulsion Laboratory Airborne Earth Science Workshop , Pasadena, CA, USA, 2004. -
- [59] N. Dobigeon and C. F´ evotte, 'Robust nonnegative matrix factorization for nonlinear unmixing of hyperspectral images,' in Proc. IEEE GRSS Workshop Hyperspectral Image SIgnal Process.: Evolution in Remote Sens. (WHISPERS) , Gainesville, FL, June 2013. -
- [54] M. Winter, 'Fast autonomous spectral end-member determination in hyperspectral data,' in Proc. 13th Int. Conf. on Applied Geologic Remote Sensing , vol. 2, Vancouver, April 1999, pp. 337-344. -
- [53] RSI (Research Systems Inc.), ENVI User's guide Version 4.0 , Boulder, CO 80301 USA, Sept. 2003. -
- [12] R. B. Singer and T. B. McCord, 'Mars: Large scale mixing of bright and dark surface materials and implications for analysis of spectral reflectance,' in Proc. 10th Lunar and Planetary Sci. Conf. , March 1979, pp. 1835-1848. -
- [49] --, 'Nonlinear abundance estimation of hyperspectral images with L1-norm spatial regularization,' IEEE Trans. Geosci. and Remote Sensing , 2013, submitted. -
Cited by
Ghamisi, P., Yokoya, N., Li, J., Liao, W., Liu, S., Plaza, J., Rasti, B., & Plaza, A. (2017) · IEEE Geoscience and Remote Sensing Magazine · doi:10.1109/mgrs.2017.2762087 · Cited by 909
“NONLINEAR UNMIXING Apart from the linear spectral unmixing algorithms, several alternatives dealing with the mixture problem can be found in the literature [166]–[168].”
BackgroundA Review of Nonlinear Hyperspectral Unmixing Methods
Heylen, R., Parente, M., & Gader, P. (2014) · IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing · doi:10.1109/jstars.2014.2320576 · Cited by 591
“This bilinear model differs from the Fan and generalized bilinear models on two accounts: selfinteractions are included, and all bilinear terms are scaled with the same constant: ∀m, k : bmk = bamak.”
MethodA Signal Processing Perspective on Hyperspectral Unmixing: Insights from Remote Sensing
Ma, W.-K., Bioucas-Dias, J. M., Chan, T.-H., Gillis, N., Gader, P., Plaza, A. J., Ambikapathi, A., & Chi, C.-Y. (2014) · IEEE Signal Processing Magazine · doi:10.1109/msp.2013.2279731 · Cited by 456
DAEN: Deep Autoencoder Networks for Hyperspectral Unmixing
Su, Y., Li, J., Plaza, A., Marinoni, A., Gamba, P., & Chakravortty, S. (2019) · IEEE Transactions on Geoscience and Remote Sensing · doi:10.1109/tgrs.2018.2890633 · Cited by 335
“Unmixing algorithms rely on specific mixing models, which can be characterized as either linear or nonlinear [7].”
BackgroundSpectral Variability in Hyperspectral Data Unmixing: A comprehensive review
Borsoi, R. A., Imbiriba, T., Bermudez, J. C. M., Richard, C., Chanussot, J., Drumetz, L., Tourneret, J.-Y., Zare, A., & Jutten, C. (2021) · IEEE Geoscience and Remote Sensing Magazine · doi:10.1109/mgrs.2021.3071158 · Cited by 265
“were originally devised as forward models that accurately describe the reflectance spectra based on a set of parameters, and were not originally designed to be inverted, which limits their use for SU in practical problems [182].”
BackgroundMethodSelf-citationHyperspectral Unmixing Using a Neural Network Autoencoder
Palsson, B., Sigurdsson, J., Sveinsson, J. R., & Ulfarsson, M. O. (2018) · IEEE Access · doi:10.1109/access.2018.2818280 · Cited by 256
“Non-linear models attempt to model the physical interactions between the light scattered by multiple materials in the scene where the interactions can be at a classical (multilayered) level, or at a microscopic (intimate) level [2], [3].”
BackgroundBlind Hyperspectral Unmixing Using an Extended Linear Mixing Model to Address Spectral Variability
Drumetz, L., Veganzones, M.-A., Henrot, S., Phlypo, R., Chanussot, J., & Jutten, C. (2016) · IEEE Transactions on Image Processing · doi:10.1109/tip.2016.2579259 · Cited by 251
“Although nonlinear unmixing has recently received much attention in the community –providing new nonlinear mixing models and the corresponding SU algorithms [8], [9]–, spectral variability has been given less attention in comparison.”
BackgroundEndNet: Sparse AutoEncoder Network for Endmember Extraction and Hyperspectral Unmixing
Ozkan, S., Kaya, B., & Akar, G. B. (2019) · IEEE Transactions on Geoscience and Remote Sensing · doi:10.1109/tgrs.2018.2856929 · Cited by 244
Hyperspectral Unmixing With Spectral Variability Using a Perturbed Linear Mixing Model
Thouvenin, P.-A., Dobigeon, N., & Tourneret, J.-Y. (2016) · IEEE Transactions on Signal Processing · doi:10.1109/tsp.2015.2486746 · Cited by 230
“different endmembers as in [32], [34], [35].”
Self-citationHyperspectral Super-Resolution: A Coupled Tensor Factorization Approach
Kanatsoulis, C. I., Fu, X., Sidiropoulos, N. D., & Ma, W.-K. (2018) · IEEE Transactions on Signal Processing · doi:10.1109/tsp.2018.2876362 · Cited by 238
“Moreover, the LMM is considered a coarse approximation of reality since factors such as endmember variability [35] and the existence of nonlinear mixtures [36] bring into question the validity of the LMM.”
He, W., Zhang, H., & Zhang, L. (2017) · IEEE Transactions on Geoscience and Remote Sensing · doi:10.1109/tgrs.2017.2683719 · Cited by 238
“The nonlinear mixing model describes the mixed spectrum by assuming that the observed pixel is generated from a nonlinear function of the abundances associated with the endmembers [10].”
BackgroundQuantum-inspired computational imaging
Altmann, Y., McLaughlin, S., Padgett, M. J., Goyal, V. K., Hero, A. O., & Faccio, D. (2018) · Science · doi:10.1126/science.aat2298 · Cited by 229
Self-citationCyCU-Net: Cycle-Consistency Unmixing Network by Learning Cascaded Autoencoders
Gao, L., Han, Z., Hong, D., Zhang, B., & Chanussot, J. (2022) · IEEE Transactions on Geoscience and Remote Sensing · doi:10.1109/tgrs.2021.3064958 · Cited by 209
“Nevertheless, the assumption of an LMM is untenable due to the existence of spectral variability (SV) [17] and nonlinear interactions [18].”
A Multilinear Mixing Model for Nonlinear Spectral Unmixing
Heylen, R., & Scheunders, P. (2016) · IEEE Transactions on Geoscience and Remote Sensing · doi:10.1109/tgrs.2015.2453915 · Cited by 215
“in [9] and [10], and detailed overviews of these models and their differences can be found in the review papers [11], [12].”
BackgroundNonlinear Hyperspectral Unmixing With Robust Nonnegative Matrix Factorization
Fevotte, C., & Dobigeon, N. (2015) · IEEE Transactions on Image Processing · doi:10.1109/tip.2015.2468177 · Cited by 210
“Conversely, as illustrated in previous works [21],”
BackgroundGuerschman, J. P., Scarth, P. F., McVicar, T. R., Renzullo, L. J., Malthus, T. J., Stewart, J. B., Rickards, J. E., & Trevithick, R. (2015) · Remote Sensing of Environment · doi:10.1016/j.rse.2015.01.021 · Cited by 200
A review on spectral processing methods for geological remote sensing
Asadzadeh, S., & de Souza Filho, C. R. (2016) · International Journal of Applied Earth Observation and Geoinformation · doi:10.1016/j.jag.2015.12.004 · Cited by 203
Spectral Unmixing of Hyperspectral Imagery Using Multilayer NMF
Rajabi, R., & Ghassemian, H. (2015) · IEEE Geoscience and Remote Sensing Letters · doi:10.1109/lgrs.2014.2325874 · Cited by 160
“There are two main classes of mixing models for spectral unmixing problem: linear mixing model (LMM) and the category of nonlinear mixing models [15].”
BackgroundFeng, X.-R., Li, H.-C., Li, J., Du, Q., Plaza, A., & Emery, W. J. (2018) · IEEE Transactions on Geoscience and Remote Sensing · doi:10.1109/tgrs.2018.2834567 · Cited by 154
Deep Generative Endmember Modeling: An Application to Unsupervised Spectral Unmixing
Borsoi, R. A., Imbiriba, T., & Bermudez, J. C. M. (2020) · IEEE Transactions on Computational Imaging · doi:10.1109/tci.2019.2948726 · Cited by 123
“Many parametric models have been proposed to describe the interaction between light and the target surface [4], [6].”
BackgroundSelf-citationLin, C.-H., Ma, F., Chi, C., & Hsieh, C.-H. (2017) · IEEE Transactions on Geoscience and Remote Sensing · doi:10.1109/tgrs.2017.2766080 · Cited by 138
“Another possible strategy is to adopt nonlinear modeling for Z [61], but it may induce a cumbersome fusion algorithm as considering multiple reflections of solar radiation can lead to far more complicated (nonconvex) formulations.”
BackgroundUnDIP: Hyperspectral Unmixing Using Deep Image Prior
Rasti, B., Koirala, B., Scheunders, P., & Ghamisi, P. (2021) · IEEE Transactions on Geoscience and Remote Sensing · doi:10.1109/tgrs.2021.3067802 · Cited by 155
“From a modeling point of view, unmixing techniques can be divided into two main groups: linear unmixing and nonlinear unmixing [3], [4].”
MethodBlind hyperspectral unmixing using an extended linear mixing model to address spectral variability
Drumetz, L., Henrot, S., Veganzones, M. A., Chanussot, J., & Jutten, C. (2015) · Workshop on Hyperspectral Image and Signal Processing: Evolution in Remote Sensing · doi:10.1109/whispers.2015.8075417 · Cited by 128
Advances in spaceborne hyperspectral remote sensing in China
Zhong, Y., Wang, X., Wang, S., & Zhang, L. (2021) · Geo-spatial Information Science · doi:10.1080/10095020.2020.1860653 · Cited by 124
Sparsity-Regularized Robust Non-Negative Matrix Factorization for Hyperspectral Unmixing
He, W. P., Zhang, H., & Zhang, L. (2016) · IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing · doi:10.1109/jstars.2016.2519498 · Cited by 120
“ear function of the abundance associated with the endmembers [5].”
BackgroundA Fast Multiscale Spatial Regularization for Sparse Hyperspectral Unmixing
Borsoi, R. A., Imbiriba, T., Bermudez, J. C. M., & Richard, C. (2018) · IEEE Geoscience and Remote Sensing Letters · doi:10.1109/lgrs.2018.2878394 · Cited by 120
“The LMM is effective in accurately modelling mixtures occurring in scenes where the materials of interest cover a large area with respect to the pixel size [2].”
InfluentialSelf-citationMiSiCNet: Minimum Simplex Convolutional Network for Deep Hyperspectral Unmixing
Rasti, B., Koirala, B., Scheunders, P., & Chanussot, J. (2022) · IEEE Transactions on Geoscience and Remote Sensing · doi:10.1109/tgrs.2022.3146904 · Cited by 133
“On the other hand, linear unmixing is found helpful in remote sensing applications, where hyperspectral images often consist of homogeneous regions of single materials [3], [4].”
BackgroundA survey on deep matrix factorizations
Handschutter, P. D., Gillis, N., & Siebert, X. (2021) · Computer Science Review · doi:10.1016/j.cosrev.2021.100423 · Cited by 128
“For example, in hyperspectral unmixing, scattering and various interactions may justify the use of non-linear models [114].”
Li, H., Feng, R., Wang, L., Zhong, Y., & Zhang, L. (2020) · IEEE Transactions on Geoscience and Remote Sensing · doi:10.1109/tgrs.2020.2994260 · Cited by 113
Generalized Linear Mixing Model Accounting for Endmember Variability
Imbiriba, T., Borsoi, R. A., & Bermudez, J. C. M. (2018) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp.2018.8462214 · Cited by 115
Self-citationBlind Hyperspectral Unmixing Using Autoencoders: A Critical Comparison
Palsson, B., Sveinsson, J. R., & Ulfarsson, M. O. (2022) · IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing · doi:10.1109/jstars.2021.3140154 · Cited by 107
“2) Nonlinear Models: When the mixing scale is not macroscopic as for intimate mixtures of materials, or the incident light scattersoffmultiplematerials,nonlinearmodelsareneeded[19], [33].”
BackgroundRobust Volume Minimization-Based Matrix Factorization for Remote Sensing and Document Clustering
Fu, X., Huang, K., Yang, B., Ma, W., & Sidiropoulos, N. D. (2016) · IEEE Transactions on Signal Processing · doi:10.1109/tsp.2016.2602800 · Cited by 105
“04 29 0v 1 [ st”
Hyperspectral Image Unmixing With Endmember Bundles and Group Sparsity Inducing Mixed Norms
Drumetz, L., Meyer, T. R., Chanussot, J., Bertozzi, A. L., & Jutten, C. (2019) · IEEE Transactions on Image Processing · doi:10.1109/tip.2019.2897254 · Cited by 104
“These physical considerations have spanned a number of approaches taking some nonlinear effects into consideration; see [9] and [10] for reviews on that subject.”
BackgroundDeep Autoencoders With Multitask Learning for Bilinear Hyperspectral Unmixing
Su, Y., Xu, X. D., Li, J., Qi, H., Gamba, P. E., & Plaza, A. (2020) · IEEE Transactions on Geoscience and Remote Sensing · doi:10.1109/tgrs.2020.3041157 · Cited by 104
“Nonlinear mixture models (NLMMs), instead, aim at characterizing the interactions between the sunlight scattered by multiple materials, being at an intimate or multilayered level [32], [33].”
BackgroundMethodInfluentialDAAN: A Deep Autoencoder-Based Augmented Network for Blind Multilinear Hyperspectral Unmixing
Su, Y., Zhu, Z., Gao, L., Plaza, A., Li, P., Sun, X., & Xu, X. D. (2024) · IEEE Transactions on Geoscience and Remote Sensing · doi:10.1109/tgrs.2024.3381632 · Cited by 102
“NSMMs contain various models that delineate the interactions of solar radiation when scattered by diverse materials [30].”
BackgroundGao, L., Wang, Z., Zhuang, L., Yu, H., Zhang, B., & Chanussot, J. (2021) · IEEE Transactions on Geoscience and Remote Sensing · doi:10.1109/tgrs.2021.3065990 · Cited by 98
“The interaction coefficients in the GBM were set randomly, and the interaction coefficients in the PPNM were set to 0.25.”
BackgroundMethodInfluentialHyperspectral Unmixing for Additive Nonlinear Models With a 3-D-CNN Autoencoder Network
Zhao, M., Wang, M., Chen, J., & Rahardja, S. (2021) · IEEE Transactions on Geoscience and Remote Sensing · doi:10.1109/tgrs.2021.3098745 · Cited by 114
Chen, J., Zhao, M., Wang, X., Richard, C., & Rahardja, S. (2023) · IEEE Signal Processing Magazine · doi:10.1109/msp.2022.3208987 · Cited by 90
“The works in [2], [5] provide thorough reviews on physical mixture models and associated unmixing methods.”
BackgroundSelf-citationOnline Dictionary Learning for Kernel LMS
Gao, W., Chen, J., Richard, C., & Huang, J. (2014) · IEEE Transactions on Signal Processing · doi:10.1109/tsp.2014.2318132 · Cited by 90
“Based on this scheme, several state-of-the-art linear methods were reconsidered to derive powerful nonlinear generalizations operating in high-dimensional RKHS [2], [3]: the recursive least-squares algorithm (RLS), the affine projection algorithm (APA), and the least-mean-square algorithm (LMS).”
Self-citationHyperspectral Unmixing Using Deep Convolutional Autoencoders in a Supervised Scenario
Khajehrayeni, F., & Ghassemian, H. (2020) · IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing · doi:10.1109/jstars.2020.2966512 · Cited by 85
“There are two types of mixing models for the HSU: The linear mixing model (LMM) and the families of the nonlinear mixing models (NMMs) [10].”
MethodAltmann, Y., Nicolas Dobigeon, B., & Tourneret, J. (2014) · IEEE Transactions on Image Processing · doi:10.1109/tip.2014.2314022 · Cited by 96
Self-citationUnsupervised Unmixing of Hyperspectral Images Accounting for Endmember Variability
Halimi, A., Dobigeon, N., & Tourneret, J. (2015) · IEEE Transactions on Image Processing · doi:10.1109/tip.2015.2471182 · Cited by 87
Self-citationA Fast Hyperplane-Based Minimum-Volume Enclosing Simplex Algorithm for Blind Hyperspectral Unmixing
Lin, C.-H., Chi, C., Wang, Y.-H., & Chan, T. (2015) · IEEE Transactions on Signal Processing · doi:10.1109/tsp.2015.2508778 · Cited by 92
“However, studying these effects is out of the scope of this paper, and the representative LMM i s sufficient for our analysis and algorithm development; interested reader s can refer to the magazine papers [34] and [35], about the non-linear effect a nd the endmember variability effect, respectively.”
BackgroundLi, S., Zhou, M., Li, Q., Hu, M., Wen, Y., Zhang, J., Lu, Y., & Chu, J. (2021) · Methods · doi:10.1016/j.ymeth.2021.04.005 · Cited by 82
Dynamical Spectral Unmixing of Multitemporal Hyperspectral Images
Henrot, S., Chanussot, J., & Jutten, C. (2016) · IEEE Transactions on Image Processing · doi:10.1109/tip.2016.2562562 · Cited by 86
Hyperspectral Unmixing in Presence of Endmember Variability, Nonlinearity, or Mismodeling Effects
Halimi, A., Honeiné, P., & Bioucas‐Dias, J. M. (2016) · IEEE Transactions on Image Processing · doi:10.1109/tip.2016.2590324 · Cited by 83
“…term of the pixel (i, j) Σ ∈ RL×L noise covariance matrix σ2 ∈ RL×1 diagonal of the noise covariance matrix M ∈ RL×R endmember matrix Si,j ∈ RL×R endmember matrix of the pixel (i, j) Y ∈ RL×N spectra of the pixels A ∈ RR×N abundance matrix ∈ R1×N variance of the residual terms (or their energy)”
Linear and Nonlinear Unmixing in Hyperspectral Imaging
Nicolas Dobigeon, B., Altmann, Y., Brun, N., & Moussaoui, S. (2016) · Data handling in science and technology · doi:10.1016/b978-0-444-63638-6.00006-1 · Cited by 98
“[154], when the nonlinear effects are not too severe and the pure pixel assumption discussed in Section 2.”
BackgroundApplication of hyperspectral remote sensing for flower mapping in African savannas
Landmann, T., Piiroinen, R., Makori, D. M., Abdel-Rahman, E. M., Makau, S., Pellikka, P., & Raina, S. K. (2015) · Remote Sensing of Environment · doi:10.1016/j.rse.2015.06.006 · Cited by 79
Robust Sparse Hyperspectral Unmixing With $\ell_{2,1}$ Norm
Ma, Y., Li, C., Mei, X., Liu, C., & Ma, J. (2016) · IEEE Transactions on Geoscience and Remote Sensing · doi:10.1109/tgrs.2016.2616161 · Cited by 78
“…the LMM is widely admitted in the geoscience and remote sensing processing areas, the LMM may be not valid in many situations, for example, when there are multiscattering effects or intimate interactions [40], and the NLMMs provide an alternative to overcoming the inherent limitations of the LMM.”
Spectral-Spatial Hyperspectral Unmixing Using Nonnegative Matrix Factorization
Zhang, S., Zhang, G., Li, F., Deng, C., Wang, S., Plaza, A., & Li, J. (2021) · IEEE Transactions on Geoscience and Remote Sensing · doi:10.1109/tgrs.2021.3074364 · Cited by 82
“The average reconstruction error [39], [40] ARE =”
BackgroundAltmann, Y., Nicolas Dobigeon, B., & Tourneret, J. (2014) · Cited by 71
Self-citationDual-Branch Subpixel-Guided Network for Hyperspectral Image Classification
Han, Z., Yang, J., Gao, L., Zeng, Z., Zhang, B., & Chanussot, J. (2024) · IEEE Transactions on Geoscience and Remote Sensing · doi:10.1109/tgrs.2024.3418583 · Cited by 84
“NLMM introduces the multiple light scatterings in the microscopic scale and achieves a suitable approximation to the actual mixing [52], [53].”
BackgroundRasti, B., Zouaoui, A., Mairal, J., & Chanussot, J. (2024) · IEEE Transactions on Geoscience and Remote Sensing · doi:10.1109/tgrs.2024.3393570 · Cited by 83
“Alternatively, nonlinear models are used [1], [2].”
BackgroundBand-Wise Nonlinear Unmixing for Hyperspectral Imagery Using an Extended Multilinear Mixing Model
Yang, B., & Wang, B. (2018) · IEEE Transactions on Geoscience and Remote Sensing · doi:10.1109/tgrs.2018.2842707 · Cited by 67
“, ground covers in flat separated areas) [4]–[7], model inaccuracies issued from nonlinear mixing effect in typical natural scenarios decrease dramatically their unmixing performances [8], [9].”
BackgroundSingle-Channel Audio Source Separation with NMF: Divergences, Constraints and Algorithms
Févotte, C., Emmanuel Vincent, A., & Ozerov, A. (2018) · Signals and communication technology · doi:10.1007/978-3-319-73031-8_1 · Cited by 64
Unsupervised Nonlinear Spectral Unmixing Based on a Multilinear Mixing Model
Wei, Q., Chen, M., Tourneret, J., & Godsill, S. (2017) · IEEE Transactions on Geoscience and Remote Sensing · doi:10.1109/tgrs.2017.2693366 · Cited by 64
“However, there exist many scenarios, involving intimate mixtures or multipath reflections, for which the LMM is not appropriate and can be advantageously replaced by a nonlinear mixture model [1], [14]–[16].”
BackgroundSelf-citationFast Hyperspectral Unmixing in Presence of Nonlinearity or Mismodeling Effects
Halimi, A., Bioucas‐Dias, J. M., Nicolas Dobigeon, B., Buller, G. S., & McLaughlin, S. (2016) · IEEE Transactions on Computational Imaging · doi:10.1109/tci.2016.2631979 · Cited by 65
“Nonlinear mixture models are an alternative to better account for those effects [12], [13] and we distinguish between two main families: the first is signal processing based and seeks to construct flexible models that can represent a wide range of nonlinearities.”
BackgroundSelf-citationHyperspectral Unmixing With Spectral Variability Using Adaptive Bundles and Double Sparsity
Uezato, T., Fauvel, M., & Dobigeon, N. (2019) · IEEE Transactions on Geoscience and Remote Sensing · doi:10.1109/tgrs.2018.2889256 · Cited by 62
Nonlinear Hyperspectral Unmixing Based on Geometric Characteristics of Bilinear Mixture Models
Yang, B., Wang, B., & Wu, Z. (2017) · IEEE Transactions on Geoscience and Remote Sensing · doi:10.1109/tgrs.2017.2753847 · Cited by 62
“In recent years, various non-LMMs and unmixing algorithms have been studied [10], [11].”
BackgroundA Novel Approach for Efficient $p$ -Linear Hyperspectral Unmixing
Marinoni, A., & Gamba, P. E. (2015) · IEEE Journal of Selected Topics in Signal Processing · doi:10.1109/jstsp.2015.2416693 · Cited by 67
“Several papers have addressed the topic of nonlinear unmixing of hyperspectral images [22].”
BackgroundUniformity-Based Superpixel Segmentation of Hyperspectral Images
Saranathan, A. M., & Parente, M. C. (2015) · IEEE Transactions on Geoscience and Remote Sensing · doi:10.1109/tgrs.2015.2480863 · Cited by 61
“In such cases, the mixing can no longer be considered linear and is modeled using nonlinear representations [2], [8], [18].”
BackgroundLow-Rank Tensor Modeling for Hyperspectral Unmixing Accounting for Spectral Variability
Imbiriba, T., Borsoi, R. A., & Bermudez, J. C. M. (2019) · IEEE Transactions on Geoscience and Remote Sensing · doi:10.1109/tgrs.2019.2949543 · Cited by 62
“Different mixing models have been employed to explain the interaction between light and the endmembers [1]-[4].”
BackgroundMethodSelf-citationA 3-D-CNN Framework for Hyperspectral Unmixing With Spectral Variability
Zhao, M., Shi, S., Chen, J., & Nicolas Dobigeon, B. (2022) · IEEE Transactions on Geoscience and Remote Sensing · doi:10.1109/tgrs.2022.3141387 · Cited by 65
“It aims to parse the mixed pixel into a set of pure spectral signatures (named endmembers) and their related fractional percentages (named abundances) [5], [6].”
BackgroundNicolas Dobigeon, B., Tits, L., Somers, B., Altmann, Y., & Coppin, P. R. (2014) · IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing · doi:10.1109/jstars.2014.2328872 · Cited by 60
“However, for specific applications, LMM has demonstrated some difficulties to accurately describe real mixtures [10].”
BackgroundChi, J., & Crawford, M. M. (2014) · IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing · doi:10.1109/jstars.2014.2319585 · Cited by 57
“combinations of the endmembers, and computing the corresponding fractions, called fractional abundances [14]–[19].”
BackgroundRomaszewski, M., Głomb, P., & Cholewa, M. (2016) · ISPRS Journal of Photogrammetry and Remote Sensing · doi:10.1016/j.isprsjprs.2016.08.011 · Cited by 56
Deep Half-Siamese Networks for Hyperspectral Unmixing
Han, Z., Hong, D., Gao, L., Zhang, B., & Chanussot, J. (2021) · IEEE Geoscience and Remote Sensing Letters · doi:10.1109/lgrs.2020.3011941 · Cited by 50
Zhu, W., Chayes, V., Tiard, A., Sanchez, S. M., Dahlberg, D., Bertozzi, A. L., Osher, S., Zosso, D. P., & Kuang, D. (2017) · IEEE Transactions on Geoscience and Remote Sensing · doi:10.1109/tgrs.2017.2654486 · Cited by 57
“Despite the simplicity of LUM, the assumption of a linear mixture of materials has been shown to be physically inaccurate in certain situations [7].”
BackgroundHierarchical Beta Process with Gaussian Process Prior for Hyperspectral Image Super Resolution
Akhtar, N., Faisal Shafait, B., & Mian, A. S. (2016) · Lecture notes in computer science · doi:10.1007/978-3-319-46487-9_7 · Cited by 55
“Gaussian Processes provide an excellent tool for modeling natural spectra [19] because they can easily incorporate the regularly occurring smoothness of spectra.”
MethodRasti, B., Koirala, B., & Scheunders, P. (2022) · IEEE Transactions on Geoscience and Remote Sensing · doi:10.1109/tgrs.2022.3202490 · Cited by 59
“In that case, nonlinear models must 50 be utilized [4], [5].”
MethodEfficient Sequential Monte-Carlo Samplers for Bayesian Inference
Nguyen, T. L. T., Septier, F., Peters, G. W., & Delignon, Y. (2015) · IEEE Transactions on Signal Processing · doi:10.1109/tsp.2015.2504342 · Cited by 60
“Although this approach presents many advantages over traditional Monte-Carlo methods, the potential of this emergent technique is however largely underexploited in signal processing.”
MethodKaroui, M. S., Benhalouche, F. Z., Deville, Y., Djerriri, K., Briottet, X., Houet, T., Bris, A. L., & Weber, C. (2019) · Remote Sensing · doi:10.3390/rs11182164 · Cited by 54
“, when the landscape is significantly non-flat and/or the irradiance is heterogeneous, especially in urban areas, LSU methods are not effective and must be replaced by nonlinear ones [12], in particular, those based on linear quadratic or bilinear mixing models [13–15].”
BackgroundAltmann, Y., Nicolas Dobigeon, B., McLaughlin, S., & Tourneret, J. (2014) · IEEE Transactions on Image Processing · doi:10.1109/tip.2014.2312616 · Cited by 55
“They have been proposed in the hyperspectral image literature and can be divided into two main classes [8].”
BackgroundSelf-citationSpectral Variability Aware Blind Hyperspectral Image Unmixing Based on Convex Geometry
Drumetz, L., Chanussot, J., Jutten, C., Ma, W., & Iwasaki, A. (2020) · IEEE Transactions on Image Processing · doi:10.1109/tip.2020.2974062 · Cited by 59
“In [14], endmember extraction algorithms are considered to be able to correctly identify endmembers in mildly nonlinearly mixed data because for many models, endmembers are still extreme points”
BackgroundRobust Linear Spectral Unmixing Using Anomaly Detection
Altmann, Y., McLaughlin, S., & Hero, A. O. (2015) · IEEE Transactions on Computational Imaging · doi:10.1109/tci.2015.2455411 · Cited by 54
“Nonlinear mixing models (NLMMs) have been proposed in the hyperspectral image literature and can be divided into two main classes [3], [4].”
BackgroundSelf-citationMei, X., Ma, Y., Li, C., Fan, F., Huang, J., & Ma, J. (2017) · Neurocomputing · doi:10.1016/j.neucom.2017.11.052 · Cited by 51
Hyperspectral unmixing for Raman spectroscopy via physics-constrained autoencoders
Georgiev, D., Fernández-Galiana, Á., Vilms Pedersen, S., Papadopoulos, G., Xie, R., Stevens, M. M., & Barahona, M. (2024) · Proceedings of the National Academy of Sciences · doi:10.1073/pnas.2407439121 · Cited by 50
“However, accounting for non-linear mixing interactions increases the complexity and computational cost of unmixing [82, 83], an issue of especial relevance in Raman spectroscopy where datasets (e.g., imaging/volumetric scans) are typically larger than in remote sensing; hence, despite its…”
BackgroundHyperspectral Unmixing Based on Multilinear Mixing Model Using Convolutional Autoencoders
Fang, T., Zhu, F., & Chen, J. (2024) · IEEE Transactions on Geoscience and Remote Sensing · doi:10.1109/tgrs.2024.3360714 · Cited by 56
“SU refers to identifying the endmembers and to estimating the corresponding fractions for each pixel, namely, abundances [1].”
BackgroundSolar photovoltaic module detection using laboratory and airborne imaging spectroscopy data
Ji, C., Bachmann, M., Esch, T., Feilhauer, H., Heiden, U., Heldens, W., Hueni, A., Lakes, T., Metz-Marconcini, A., Schroedter‐Homscheidt, M., et al. (2021) · Remote Sensing of Environment · doi:10.1016/j.rse.2021.112692 · Cited by 52
AutoNAS: Automatic Neural Architecture Search for Hyperspectral Unmixing
Han, Z., Hong, D., Gao, L., Zhang, B., Huang, M. J., & Chanussot, J. (2022) · IEEE Transactions on Geoscience and Remote Sensing · doi:10.1109/tgrs.2022.3186480 · Cited by 54
Blind Hyperspectral Unmixing Using Total Variation and $\ell_q$ Sparse Regularization
Sigurðsson, J., Ulfarsson, M. O., & Sveinsson, J. R. (2016) · IEEE Transactions on Geoscience and Remote Sensing · doi:10.1109/tgrs.2016.2582824 · Cited by 51
“For a recent review of nonlinear unmixing methods, we refer the reader to [4] and [5].”
MethodPatel, A. K., Ghosh, J. K., Pande, S., & Sayyad, S. U. (2020) · IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing · doi:10.1109/jstars.2020.3039844 · Cited by 50
SRViT: Self-Supervised Relation-Aware Vision Transformer for Hyperspectral Unmixing
Su, Y., Gao, L., Plaza, A., Sun, X., Jiang, M., & Yang, G. (2025) · IEEE Transactions on Neural Networks and Learning Systems · doi:10.1109/tnnls.2025.3571798 · Cited by 51
“Abun-dance maps indicate the relative content of individual materials in each pixel, usually expressed as fraction values [6].”
A Data Dependent Multiscale Model for Hyperspectral Unmixing With Spectral Variability
Borsoi, R. A., Imbiriba, T., & Bermudez, J. C. M. (2020) · IEEE Transactions on Image Processing · doi:10.1109/tip.2020.2963959 · Cited by 50
“Although this assumption naturally leads to fast and reliable unmixing strategies, the intrinsic limitation of the LMM cannot cope with relevant nonideal effects intrinsic to practical applications [3], [5], [6].”
BackgroundSelf-citationHyperspectral Unmixing with Robust Collaborative Sparse Regression
Li, C., Ma, Y., Mei, X., Liu, C., & Ma, J. (2016) · Remote Sensing · doi:10.3390/rs8070588 · Cited by 46
“NLMMs have been proposed in the hyperspectral image processing and can be divided into two main classes [21].”
BackgroundCalibration and validation of the Australian fractional cover product for MODIS collection 6
Guerschman, J. P., & Hill, M. J. (2018) · Remote Sensing Letters · doi:10.1080/2150704x.2018.1465611 · Cited by 45
“In particular, this model inverts the field datasets and derives endmembers by regressing the reflectance values against the field data (Dobigeon et al. 2014; Puyou-Lascassies et al. 1994).”
MethodAn Abundance Characteristic-Based Independent Component Analysis for Hyperspectral Unmixing
Wang, N., Du, B., Zhang, L., & Zhang, L. (2014) · IEEE Transactions on Geoscience and Remote Sensing · doi:10.1109/tgrs.2014.2322862 · Cited by 46
“For an overview of recent advances in nonlinear unmixing, see [6].”
BackgroundHyperspectral Image Unmixing With LiDAR Data-Aided Spatial Regularization
Uezato, T., Fauvel, M., & Nicolas Dobigeon, B. (2018) · IEEE Transactions on Geoscience and Remote Sensing · doi:10.1109/tgrs.2018.2823419 · Cited by 46
“consequence, these DSM-based weighted regularizations can be coupled with the various mixing models incorporating linear or nonlinear interaction [39] or spectral variability [41].”
BackgroundNonlinear Hyperspectral Unmixing Using Nonlinearity Order Estimation and Polytope Decomposition
Marinoni, A., Plaza, J., Plaza, A., & Gamba, P. E. (2015) · IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing · doi:10.1109/jstars.2015.2427517 · Cited by 43
“Indeed, as photonic interactions or multilayer mixtures show up, NLMMs that thoroughly describe the intimate mixtures of the constituent materials have been proposed [7].”
BackgroundMethodModel-Based Deep Autoencoder Networks for Nonlinear Hyperspectral Unmixing
Li, H., Borsoi, R. A., Imbiriba, T., Closas, P., Bermudez, J. C. M., & Deniz Erdogmus, B. (2021) · IEEE Geoscience and Remote Sensing Letters · doi:10.1109/lgrs.2021.3075138 · Cited by 43
“Prominent examples include the BLMM (for macroscopic interactions), the PNMM (e.g., nonlinearities occurring between the scene and the sensor), and Hapke (for intimate mixtures) models [1].”
InfluentialSelf-citationBiobjective Nonnegative Matrix Factorization: Linear Versus Kernel-Based Models
Zhu, F., & Honeiné, P. (2016) · IEEE Transactions on Geoscience and Remote Sensing · doi:10.1109/tgrs.2016.2535298 · Cited by 44
Estimating the Intrinsic Dimension of Hyperspectral Images Using a Noise-Whitened Eigengap Approach
Halimi, A., Honeiné, P., Kharouf, M., Richard, C., & Tourneret, J. (2016) · IEEE Transactions on Geoscience and Remote Sensing · doi:10.1109/tgrs.2016.2528298 · Cited by 42
“[3]–[5], the linear mixing model is widely accepted because of its natural physical interpretation.”
BackgroundSelf-citationHarmonic Mixture Modeling for Efficient Nonlinear Hyperspectral Unmixing
Marinoni, A., Plaza, A., & Gamba, P. E. (2016) · IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing · doi:10.1109/jstars.2016.2514740 · Cited by 41
“However, as the recorded scenes can be very complex both geometrically and spectrally, nonlinear interactions are crucial to assess the microscopic and/or macroscopic interactions among materials in the scene [3], [4].”
BackgroundArrigoni, S., Turra, G., & Signoroni, A. (2017) · Computers in Biology and Medicine · doi:10.1016/j.compbiomed.2017.06.018 · Cited by 46
Hyperspectral Super-Resolution via Global–Local Low-Rank Matrix Estimation
Wu, R., Ma, W.-K., Fu, X., & Li, Q. (2020) · IEEE Transactions on Geoscience and Remote Sensing · doi:10.1109/tgrs.2020.2979908 · Cited by 41
“The lowdimensional model may also be applicable to some classes of nonlinear spectral mixtures such as the bilinear mixture model [2], [3].”
MethodAdvancement of hyperspectral image processing and information extraction
Bing, Z. (2016) · National Remote Sensing Bulletin · doi:10.11834/jrs.20166179 · Cited by 50
Nonlinear Unmixing of Hyperspectral Data With Vector-Valued Kernel Functions
Ammanouil, R., Ferrari, A., Richard, C., & Mathieu, S. (2016) · IEEE Transactions on Image Processing · doi:10.1109/tip.2016.2627815 · Cited by 46
“More recently, there has been a considerable amount of studies devoted to nonlinear mixing models [8], [9].”
BackgroundSelf-citationAltmann, Y., Pereyra, M., & Bioucas‐Dias, J. M. (2015) · IEEE Transactions on Image Processing · doi:10.1109/tip.2015.2487862 · Cited by 40
“2487862 attention over the last few years [1]–[3].”
BackgroundOne-Step Generalized Likelihood Ratio Test for Subpixel Target Detection in Hyperspectral Imaging
Vincent, F., & Besson, O. (2020) · IEEE Transactions on Geoscience and Remote Sensing · doi:10.1109/tgrs.2020.2965212 · Cited by 39
“This simple and widely used model does not consider multiple light reflections between these different components, which can lead to more complicated non linear mixing models (NLMMs) [9].”
MethodA graph Laplacian regularization for hyperspectral data unmixing
Ammanouil, R., Ferrari, A., & Richard, C. (2015) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp.2015.7178248 · Cited by 39
“While there are also works in the literature that consider nonlinear mixing models [2, 3], most unmixing methods focus on the Linear Mixing model [4], where each pixel is modeled by a convex combination of the endmembers weighted by their abundances.”
BackgroundSelf-citation
Metadata
- Authors
- Dobigeon, N., Tourneret, J.-Y., Richard, C., Bermudez, J. C. M., McLaughlin, S., & Hero, A. O.
- Affiliations
- Institut National Polytechnique de Toulouse (FR)
- Université de Toulouse (FR)
- Universidade Federal de Santa Catarina (BR)
- Heriot-Watt University (GB)
- University of Michigan (US)
- Venue
- IEEE Signal Processing Magazine
- Published
- 2014-01
- Volume
- 31
- Issue
- 1
- Pages
- 82-94
- ISSN
- 1053-5888 (ISSN-L)
- OpenAlex
- W2005818237
- Cited by
- 420
- References
- 57
- Citations by year
- 2026: 11 · 2025: 23 · 2024: 32 · 2023: 38 · 2022: 36 · 2021: 34 · 2020: 45 · 2019: 31 · 2018: 36 · 2017: 39 · 2016: 45 · 2015: 29 · 2014: 20
- FWCI
- 61.72
- Type
- journal article
- Language
- en
- Open access
- green
- arxiv.org
- Locations
- IEEE Signal Processing Magazine (publishedVersion)
- arXiv (Cornell University) (submittedVersion)
- HAL (submittedVersion, other-oa)
- arXiv (Cornell University)
- Funders
- Agence Nationale de la Recherche (ANR-12-BS03-0003)
- Engineering and Physical Sciences Research Council (EP/J015180/1)
- Topics
- Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Remote-Sensing Image Classification
- Physical Sciences › Chemistry › Analytical Chemistry › Spectroscopy and Chemometric Analyses
- Physical Sciences › Earth and Planetary Sciences › Geochemistry and Petrology › Geochemistry and Geologic Mapping
- Keywords
- nonlinear unmixing, hyperspectral unmixing, linear mixing model, multiple scattering, end-member mixing model, endmember extraction
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.