Hyperspectral Unmixing Overview: Geometrical, Statistical, and Sparse Regression-Based Approaches
By José M. Bioucas-Dias, Antonio J. Plaza, Nicolas Dobigeon, Mario Parente, Qian Du, Paul D. Gader, Jocelyn Chanussot. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2012
2,769 citations recorded by the source corpus.
Imaging spectrometers measure electromagnetic energy scattered in their instantaneous field view in hundreds or thousands of spectral channels with higher spectral resolution than multispectral cameras. Imaging spectrometers are therefore often referred to as hyperspectral cameras (HSCs). Higher spectral resolution enables material identification via spectroscopic analysis, which facilitates countless applications that require identifying materials in scenarios unsuitable for classical spectroscopic analysis. Due to low spatial resolution of HSCs, microscopic material mixing, and multiple scattering, spectra measured by HSCs are mixtures of spectra of materials in a scene. Thus, accurate estimation requires unmixing. Pixels are assumed to be mixtures of a few materials, called endmembers. Unmixing involves estimating all or some of: the number of endmembers, their spectral signatures, and their abundances at each pixel. Unmixing is a challenging, ill-posed inverse problem because of model inaccuracies, observation noise, environmental conditions, endmember variability, and data set size. Researchers have devised and investigated many models searching for robust, stable, tractable, and accurate unmixing algorithms. This paper presents an overview of unmixing methods from the time of Keshava and Mustard's unmixing tutorial to the present. Mixing models are first discussed. Signal-subspace, geometrical, statistical, sparsity-based, and spatial-contextual unmixing algorithms are described. Mathematical problems and potential solutions are described. Algorithm characteristics are illustrated experimentally.
Subfields: Analytical Chemistry, Artificial Intelligence, Media Technology
View Paper: https://www.semanticscholar.org/paper/b830ba166a6c56706ae6299792f0894e4f36ca12
How researchers see this paper
The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.
In Analytical Chemistry
- Antonio J. Plaza
- Jon Atli Benediktsson
- Jocelyn Chanussot
- José M. Bioucas-Dias
- Jun Yu Li
- Pedram Ghamisi
- Nicolas Gillis
- Shutao Li
In Artificial Intelligence
- José M. Bioucas-Dias
- Jocelyn Chanussot
- Danfeng Hong
- Antonio J. Plaza
- Nicolas Gillis
- Jing Yao
- Qian Du
- Bing Zhang
In Media Technology
- Jocelyn Chanussot
- Shutao Li
- Antonio J. Plaza
- Leyuan Fang
- José M. Bioucas-Dias
- Jon Atli Benediktsson
- Pedram Ghamisi
- Jun Yu Li
Citations in the corpus
- Hyperspectral Imaging: A Review on UAV-Based Sensors, Data Processing and Applications for Agriculture and Forestry (2017, background): “According to Bioucas-Dias [54] spectral unmixing (see, for example, [74,75]) purpose is to restore the ability of ascertain materials, i.”
- Collaborative Sparse Regression for Hyperspectral Unmixing (2014, background): “We consider only scenes affected by noise, because in the noiseless case the true solutions can always be recovered with very high accuracy [2].”
- A Review of Nonlinear Hyperspectral Unmixing Methods (2014, background): “Another problem aggravating the spectral unmixing problem is that the number of endmembers M is usually unknown, and has to be estimated by some means.”
- Deep learning classifiers for hyperspectral imaging: A review (2019, background): “The most popular ones include: (i) spectral unmixing (Bioucas-Dias et al., 2012; Heylen et al., 2014; Shi and Wang, 2014; Sánchez et al., 2015; Zhong et al., 2016a), (ii) resolution enhancement (Eismann and Hardie, 2005; Mookambiga and Gomathi, 2016; Yi et al.”
- Hyperspectral Image Classification—Traditional to Deep Models: A Survey for Future Prospects (2021, background): “Thus, considering the aforementioned limitations, HSI analysis is categorized into the following main streams: dimensionality reduction [6]–[8], spectral unmixing [9]–[14], change detection [15]–[17] classification [18]–[21], feature learning for classification [22]–[24], restoration and denoising [”
- Hyperspectral Image Denoising Employing a Spatial–Spectral Deep Residual Convolutional Neural Network (2018, background): “Hyperspectral images (HSIs), which simultaneously acquire both spatial and spectral information, have already been applied in many remote sensing applications, such as classification [1], target detection [2], unmixing [3], etc.”
- Hyperspectral Remote Sensing Data Analysis and Future Challenges (2013, methodology): “6, are the linear model, the bilinear model, and the Hapke model [3], [40].”
- Fast Hyperspectral Image Denoising and Inpainting Based on Low-Rank and Sparse Representations (2018, methodology): “This paper explores two main characteristics of hyper-spectral data: 1) HSIs live in low dimensional subspaces, which opens the door to remove the bulk of the noise using projection-based methods [1]; 2) the eigen-images are self-similar and, therefore, they may be denoised with non-local patch-base”
- A Signal Processing Perspective on Hyperspectral Unmixing: Insights from Remote Sensing (2014, background): “There is a large amount of experimental evidence that indicates that nonnegative 1 , relaxation can yield useful unmixing results [2], [50], [51].”
- Advances in Hyperspectral Image and Signal Processing: A Comprehensive Overview of the State of the Art (2017, background): “Another group of abundance estimation approaches based on blind-source separation, which does not require endmember signatures to be known a priori, has been developed [134].”
- Nonlinear Unmixing of Hyperspectral Images: Models and Algorithms (2013, methodology): “In [1], it is explained that linear mixtures are reasonable when two assumptions are wholly fulfilled.”
- Total-Variation-Regularized Low-Rank Matrix Factorization for Hyperspectral Image Restoration (2016, background): “fication [1], unmixing [2], [3], and target detection [4].”
- Alternating direction algorithms for constrained sparse regression: Application to hyperspectral unmixing (2010, neutral): “Hyperspectral unmixing (HU) is a source separation problem with applications in remote sensing, analytical chemistry, and other areas [2, 11, 12, 13].”