A Convex Formulation for Hyperspectral Image Superresolution via Subspace-Based Regularization

By Miguel Simões, José M. Bioucas-Dias, Luís B. Almeida, Jocelyn Chanussot. IEEE Transactions on Geoscience and Remote Sensing, 2014

761 citations recorded by the source corpus.

Hyperspectral remote sensing images (HSIs) usually have high spectral resolution and low spatial resolution. Conversely, multispectral images (MSIs) usually have low spectral and high spatial resolutions. The problem of inferring images that combine the high spectral and high spatial resolutions of HSIs and MSIs, respectively, is a data fusion problem that has been the focus of recent active research due to the increasing availability of HSIs and MSIs retrieved from the same geographical area. We formulate this problem as the minimization of a convex objective function containing two quadratic data-fitting terms and an edge-preserving regularizer. The data-fitting terms account for blur, different resolutions, and additive noise. The regularizer, a form of vector total variation, promotes piecewise-smooth solutions with discontinuities aligned across the hyperspectral bands. The downsampling operator accounting for the different spatial resolutions, the nonquadratic and nonsmooth nature of the regularizer, and the very large size of the HSI to be estimated lead to a hard optimization problem. We deal with these difficulties by exploiting the fact that HSIs generally “live” in a low-dimensional subspace and by tailoring the split augmented Lagrangian shrinkage algorithm (SALSA), which is an instance of the alternating direction method of multipliers (ADMM), to this optimization problem, by means of a convenient variable splitting. The spatial blur and the spectral linear operators linked, respectively, with the HSI and MSI acquisition processes are also estimated, and we obtain an effective algorithm that outperforms the state of the art, as illustrated in a series of experiments with simulated and real-life data.

Subfields: Computer Vision and Pattern Recognition, Media Technology

View Paper: https://www.semanticscholar.org/paper/aac4eaeedc1283f08ef9a50117700fd038d9aac7

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computer Vision and Pattern Recognition

In Media Technology

Citations in the corpus

Fusing Hyperspectral and Multispectral Images via Coupled Sparse Tensor Factorization (2018, methodology): “…method with the following four current state-of-the-art HSI-MSI fusion methods: GSOMP (generalization of simultaneous orthogonal matching pursuit), [24], HySure (subspace regularization) [26], BSR (Bayesian sparse representation) [19], and NLSTF (non-local sparse tensor factorization) [30].”
Hyperspectral Image Super-Resolution via Non-local Sparse Tensor Factorization (2017, methodology): “[24] use a total variation regularizer to favour spatial smoothness of the solution.”
Hyperspectral Super-Resolution by Coupled Spectral Unmixing (2015, methodology): “Interestingly, among the baselines [21, 27] perform clearly worse than the others in terms of SAM, but somewhat better in terms of RMSE.”
Hyperspectral Pansharpening: A Review (2015, methodology): “3) HySure: The works [35], [54] introduce a convex regularization problem which can be seen under a Bayesian framework.”
Multisource and Multitemporal Data Fusion in Remote Sensing: A Comprehensive Review of the State of the Art (2019, neutral): “[67] M.”
Hyperspectral and Multispectral Data Fusion: A comparative review of the recent literature (2017, methodology): “This idea of fusing the HS-MS images based on spectral information of both input images on a subspace has been the main source of inspiration for many HS-MS fusion methods developed later [37]–[39].”
Fast Fusion of Multi-Band Images Based on Solving a Sylvester Equation (2015, methodology): “This section applies the proposed fusion method to three kinds of priors that have been investigated in [23], [20] and [24] for the fusion of multi-band images.”
Deep Hyperspectral Image Sharpening (2018, positive): “To better use the prior information of the HR-HSI, approaches proposed in [1], [39], and [40] also take the spatial structures of the HR-HSI into consideration to regularize the fusion problem.”
Learning a Low Tensor-Train Rank Representation for Hyperspectral Image Super-Resolution (2019, methodology): “[31] use the total variation to promote the spatial smoothness.”