Hyperspectral (HS) imaging, also known as image spectrometry, is a landmark technique in geoscience and remote sensing (RS). In the past decade, enormous efforts have been made to process and analyze these HS products, mainly by seasoned experts. However, with an ever-growing volume of data, the bulk of costs in manpower and material resources poses new challenges for reducing the burden of manual labor and improving efficiency. For this reason, it is urgent that more intelligent and automatic approaches for various HS RS applications be developed. Machine learning (ML) tools with convex optimization have successfully undertaken the tasks of numerous artificial intelligence (AI)-related applications; however, their ability to handle complex practical problems remains limited, particularly for HS data, due to the effects of various spectral variabilities in the process of HS imaging and the complexity and redundancy of higher-dimensional HS signals. Compared to convex models, nonconvex modeling, which is capable of characterizing more complex real scenes and providing model interpretability technically and theoretically, has proven to be a feasible solution that reduces the gap between challenging HS vision tasks and currently advanced intelligent data processing models.

In China, the demand for a more precise perception of the national land surface has become most urgent given the pace of development and urbanization. Constructing a very-high-resolution (VHR) land-cover dataset for China with national coverage, however, is a nontrivial task. Thus, this has become an active area of research that is impeded by the challenges of image acquisition, manual annotation, and computational complexity. To fill this gap, the first 1 m resolution national-scale land-cover map of China, SinoLC-1, was established using a deep-learning-based framework and open-access data, including global land-cover (GLC) products, OpenStreetMap (OSM), and Google Earth imagery. Reliable training labels were generated by combining three 10 m GLC products and OSM data. These training labels and 1 m resolution images derived from Google Earth were used to train the proposed framework. This framework resolved the label noise stemming from a resolution mismatch between images and labels by combining a resolution-preserving backbone, a weakly supervised module, and a self-supervised loss function, to refine the VHR land-cover results automatically without any manual annotation requirement. Based on large-storage and computing servers, processing the 73.25 TB dataset to obtain the SinoLC-1 covering the entirety of China, ∼ 9 600 000 km², took about 10 months. The SinoLC-1 product was validated using a visually interpreted validation set including over 100 000 random samples and a statistical validation set collected from the official land survey report provided by the Chinese government. The validation results showed that SinoLC-1 achieved an overall accuracy of 73.61 % and a κ coefficient of 0.6595. Validations for every provincial region further indicated the accuracy of this dataset across the whole of China. Furthermore, the statistical validation results indicated that SinoLC-1 conformed to the official survey reports with an overall misestimation rate of 6.4 %. In addition, SinoLC-1 was compared with five other widely used GLC products. These results indicated that SinoLC-1 had the highest spatial resolution and the finest landscape details. In conclusion, as the first 1 m resolution national-scale land-cover map of China, SinoLC-1 delivered accuracy and provided primal support for related research and applications throughout China. The SinoLC-1 land-cover product is freely accessible at https://doi.org/10.5281/zenodo.7707461 (Li et al., 2023).

Hyperspectral images (HSIs) are usually contaminated by various kinds of noise, such as stripes, deadlines, impulse noise, Gaussian noise, and so on, which significantly limits their subsequent application. In this paper, we model the stripes, deadlines, and impulse noise as sparse noise, and propose a unified mixed Gaussian noise and sparse noise removal framework named spatial–spectral total variation regularized local low-rank matrix recovery (LLRSSTV). The HSI is first divided into local overlapping patches, and rank-constrained low-rank matrix recovery is adopted to effectively separate the low-rank clean HSI patches from the sparse noise. Differing from the previous low-rank-based HSI denoising approaches, which process all the patches individually, a global spatial–spectral total variation regularized image reconstruction strategy is utilized to ensure the global spatial–spectral smoothness of the reconstructed image from the low-rank patches. In return, the globally reconstructed HSI further promotes the separation of the local low-rank components from the sparse noise. An augmented Lagrange multiplier method is adopted to solve the proposed LLRSSTV model, which simultaneously explores both the local low-rank property and the global spatial–spectral smoothness of the HSI. Both simulated and real HSI experiments were conducted to illustrate the advantage of the proposed method in HSI denoising, from visual/quantitative evaluations and time cost.

Non-local low-rank tensor approximation has been developed as a state-of-the-art method for hyperspectral image (HSI) restoration, which includes the tasks of denoising, compressed HSI reconstruction and inpainting. Unfortunately, while its restoration performance benefits from more spectral bands, its runtime also substantially increases. In this paper, we claim that the HSI lies in a global spectral low-rank subspace, and the spectral subspaces of each full band patch group should lie in this global low-rank subspace. This motivates us to propose a unified paradigm combining the spatial and spectral properties for HSI restoration. The proposed paradigm enjoys performance superiority from the non-local spatial denoising and light computation complexity from the low-rank orthogonal basis exploration. An efficient alternating minimization algorithm with rank adaptation is developed. It is done by first solving a fidelity term-related problem for the update of a latent input image, and then learning a low-dimensional orthogonal basis and the related reduced image from the latent input image. Subsequently, non-local low-rank denoising is developed to refine the reduced image and orthogonal basis iteratively. Finally, the experiments on HSI denoising, compressed reconstruction, and inpainting tasks, with both simulated and real datasets, demonstrate its superiority with respect to state-of-the-art HSI restoration methods.

Blind hyperspectral unmixing (HU), which includes the estimation of endmembers and their corresponding fractional abundances, is an important task for hyperspectral analysis. Recently, nonnegative matrix factorization (NMF) and its extensions have been widely used in HU. Unfortunately, most of the NMF-based methods can easily lead to an unsuitable solution, due to the nonconvexity of the NMF model and the influence of noise. To overcome this limitation, we make the best use of the structure of the abundance maps, and propose a new blind HU method named total variation regularized reweighted sparse NMF (TV-RSNMF). First, the abundance matrix is assumed to be sparse, and a weighted sparse regularizer is incorporated into the NMF model. The weights of the weighted sparse regularizer are adaptively updated related to the abundance matrix. Second, the abundance map corresponding to a single fixed endmember should be piecewise smooth. Therefore, the TV regularizer is adopted to capture the piecewise smooth structure of each abundance map. In our multiplicative iterative solution to the proposed TV-RSNMF model, the TV regularizer can be regarded as an abundance map denoising procedure, which improves the robustness of TV-RSNMF to noise. A number of experiments were conducted in both simulated and real-data conditions to illustrate the advantage of the proposed TV-RSNMF method for blind HU.

Mixed noise (such as Gaussian, impulse, stripe, and deadline noises) contamination is a common phenomenon in hyperspectral imagery (HSI), greatly degrading visual quality and affecting subsequent processing accuracy. By encoding sparse prior to the spatial or spectral difference images, total variation (TV) regularization is an efficient tool for removing the noises. However, the previous TV term cannot maintain the shared group sparsity pattern of the spatial difference images of different spectral bands. To address this issue, this article proposes a group sparsity regularization of the spatial difference images for HSI restoration. Instead of using ℓ₁- or ℓ₂-norm (sparsity) on the difference image itself, we introduce a weighted ℓ2,1-norm to constrain the spatial difference image cube, efficiently exploring the shared group sparse pattern. Moreover, we employ the well-known low-rank Tucker decomposition to capture the global spatial-spectral correlation from three HSI dimensions. To summarize, a weighted group sparsity-regularized low-rank tensor decomposition (LRTDGS) method is presented for HSI restoration. An efficient augmented Lagrange multiplier algorithm is employed to solve the LRTDGS model. The superiority of this method for HSI restoration is demonstrated by a series of experimental results from both simulated and real data, as compared with the other state-of-the-art TV-regularized low-rank matrix/tensor decomposition methods.