Hyperspectral image (HSI) super-resolution reconstruction is a challenging ill-posed inverse problem, which seeks to enhance the spatial resolution of low-resolution hyperspectral images (LR-HSIs) by integrating complementary information from high-resolution multispectral images (HR-MSIs), ultimately generating high-resolution HSIs (HR-HSIs). Existing methods commonly employ residual connections and deep layer stacking to facilitate information propagation. While residual connections effectively preserve gradient flow, we observe that naively increasing network depth in high-dimensional spectral tasks can lead to feature redundancy and performance saturation. To address these challenges, this article presents a novel Butterfly residual network (BRNet) that incorporates spectral Transformers and depth-wise convolutions to optimize both accuracy and computational efficiency of hyperspectral super-resolution reconstruction from two perspectives: learning strategy and feature extraction. Regarding learning strategy, a recursive structure coupled with a fusion parameter generation technique is proposed to promote efficient feature fusion and enable adaptive network pruning, thereby reducing redundant information and enhancing computational efficiency. For feature extraction, spectral Transformer and depth-wise convolution are employed to capture spectral and spatial features, respectively, effectively leveraging their complementary advantages across different dimensions. A specialized spectral-spatial interaction (SSI) module is then incorporated to effectively fuse the extracted features, thereby enriching the diversity of network features. Additionally, the convolutional gated feed-forward network (FFN) is designed to bolster the network's ability to capture local features while significantly reducing the computational complexity. Experimental evaluations on three hyperspectral datasets demonstrate that the proposed method outperforms existing state-of-the-art super-resolution reconstruction methods across various performance metrics, validating its effectiveness and superiority.

Land surface temperature (LST) is a crucial variable in the Earth’s surface system, playing a key role in understanding the exchanges of material and energy between the surface and the atmosphere. Hyperspectral thermal infrared (TIR) data provide new opportunities for developing methods to retrieve LST from satellite observations. However, the typical physical hyperspectral TIR LST retrieval methods are limited by their reliance on accurate atmospheric correction and specific assumptions, which would introduce complexity and reduce applicability. To address these challenges, this study presents a novel LST retrieval framework that combines a Deep Residual Regression Network (DR2N) with the Iterative Spectrally Smooth Temperature and Emissivity Separation (ISSTES) algorithm, refined by Hampel filtering. In this framework, DR2N is trained on simulated data to efficiently retrieve atmospheric parameters, including upwelling radiance, downwelling radiance and transmissivity, and after that the refined ISSTES algorithm is applied to simulated data covering various surface types, yielding an overall RMSE of 1.89 K and a bias of -0.19 K. Subsequently, to further verify the performance of the proposed algorithm, LSTs over four study areas-Spain, North Africa, Hulunbuir, and the Yellow Sea are retrieved, and are compared with IASI L2 surface temperature products originating from the European Organization for the Exploitation of Meteorological Satellites (EUMETSAT), showing an RMSE of 0.51 K and a bias of -0.25 K. This confirms the efficacy of the proposed LST retrieval approach.

Land surface temperature (LST) is a critical indicator of the interactions between the Earth’s surface and atmosphere and has long been available from satellite observations in the thermal infrared (TIR) region. Recognized as a primary way to evaluate the accuracy of LSTs, in situ validation is still a challenging task because of uncertainties in ground measurements, spatial scale mismatch between ground and satellite-based measurements, the heterogeneity of natural land surfaces, etc., leading to a lack of consistency among sets of validation results; therefore, to improve robustness against uncertainties, an optimized approach for LST validation via the Kalman filter is presented, and prediction of comprehensive validation estimate (CVE) which is close to “true” value, and more precise than those based on a single measurement alone is obtained. After the uncertainties involved in the validation process are constrained, this method is applied to FengYun-3D (FY-3D)/Medium Resolution Spectral Imager II (MERSI-II) LSTs with ground measurements from four sites in China. The results indicate that the CVE is 1.11 K, with an uncertainty of 0.07 K. Additionally, a comparison is performed with the weighted average method, and the efficacy of the Kalman filter approach in enhancing the validation accuracy is confirmed.

Self-attention-based approaches that leverage global context information for hyperspectral image (HSI) classification have gained increasing prominence. Nevertheless, due to the assignment of equivalent attention weight to all the tokens (pixels or patches), the existing self-attention mechanism inadvertently prioritizes the non-label-specified information over the instinct label-specified information, which generates attention shifts and redundancy in HSI classification. To alleviate the mentioned barrier, we propose the center-specific perception transformer network (CP-Transformer), which is the first attempt to perform class-guided attention and filter interference factors for HSI classification feature representation. Specifically, the central-pixel focus attention module (CFA) is presented to compute the label-related attention between the center and other pixels. In this manner, CFA reduces computational complexity and closely aligns with the center-pixel labeling strategy. Besides, the spectral saliency focus attention module (SSFA) is developed to capture the spectral correlation by focusing salient bands to provide a beneficial supplement for spatial features. Moreover, the hierarchical integration network (HIN) constructs the inference network to integrate and rectify spatial-spectral features for HSI classification. The experiment results on four popular HSI datasets demonstrate that the proposed method achieves robust performance compared to other state-of-the-art methods. Our code will be released at https://github.com/Chirsycy/CP-Transformer.

Due to limitations of satellite imaging systems, hyperspectral image (HSI) often suffers from incomplete coverage, with certain regions of the study area missing. Data fusion and reconstruction are effective approaches to resolve the contradiction in spatial and spectral domains, where related theories have intensively developed in recent years. However, existing fusion methods are mostly applicable to simulated data and are challenging to apply to real data. In this paper, we propose an unsupervised fusion spatial reconstruction network namely UFSRnet, which not only reconstructs the missing regions of HSI but also learns the differences between heterogeneous data adaptively. Specifically, a sensor radiation deviation correction (SRDC) module is designed to tackle the disparities between heterogeneous data adaptively. The model demonstrates commendable performance across both simulated and real data sets.

Recently, the methods based on self-attention mechanisms have gained increasing prominence in hyperspectral image classification (HSIC). However, the existing self-attention mechanism suffers the challenge of attention shift and redundancy. To address the problem, we propose the center category focusing transformer network (CCSF-Transformer) for HSIC, which is designed to resolve attention shifts and redundancy by balancing the multiple category features. Specifically, the central-category-focused attention mechanism (CFA) is presented in the proposed framework to compute the category-matched attention between the center pixel and neighbor pixels, closely matching the center-pixel style labeling strategy, and reducing the computation complexity by excluding the computation between interference pixels. Besides, the spectral-salient-focused attention module (SFA) is developed to capture the spectral correlation, which concentrates on the salient bands and suppresses the expression of redundant bands. Moreover, the hierarchical integration network (HIN) is built to rectify spatial and spectral features The experiment results on two popular HSI datasets demonstrate that the proposed method achieves robust performance compared to other state-of-the-art methods.

Surface temperature (ST) is a vital physical parameter influencing surface-atmosphere interactions. This study presents an uncertainty-based validation approach applied to Sentinel-3/SLSTR land surface temperature (LST) and sea surface temperature (SST) products.In situmeasurements were obtained from various sites in China, namely, the Dunhuang Gobi site (DHGS), Huailai Guanting Reservoir site (HGRS), Wuliangsuhai Lake site (WLSLS) and Yantai Ocean site (YTOS). The spatial representativeness ofin situmeasurements at each site was assessed using available clear-sky and high-quality ASTER LST products from April 2000 to June 2023. The four sites exhibited high spatial homogeneity, demonstrating suitability for validating STs. Therefore,in situmeasurements from these homogeneous sites were used to validate the Sentinel-3/SLSTR ST products during the daytime and nighttime using a temperature-based method. The results showed that the root mean square error (RMSE) values are lower than 1.6 K, except for those at DHGS. Furthermore, since ground-based ST validation is affected by the coupled effects of surface and atmospheric characteristics, the validation results are different under different atmospheric and surface conditions. Consequently, assessing the consistency among multiple validation results becomes challenging. To address this issue, by assuming the independence of the validation samples, we propose a method for obtaining the key comparison reference value (KCRV) from multiple validation results based on Sentinel-3/SLSTR ST products. The KCRV is close to the ‘true’ value, indicating the high quality of the validation results. For the Sentinel-3A/SLSTR and Sentinel-3B/SLSTR LST products, the KCRVs are 1.91 K and 1.71 K, respectively, with corresponding uncertainties of 0.08 K and 0.08 K, respectively. Similarly, for the Sentinel-3A/SLSTR and Sentinel-3B/SLSTR SST products, the KCRVs are 0.78 K and 0.71 K, respectively, with uncertainties of 0.08 K and 0.07 K, respectively.

Thermal infrared hyperspectral imagery presents a superior capability for capturing intricate spectral details of atmospheres and ground objects compared to multispectral images, thus offering a more nuanced dataset for land surface temperature (LST) retrieval. However, extensive inter-band correlations pose computational challenges and undesirable “dimension disaster” problem. To address this issue, this paper proposes a purpose-built framework of thermal infrared hyperspectral band selection using graph neural network for LST retrieval. Specifically, the thermal infrared hyperspectral data is firstly mapped onto a graph topology, followed by feeding it into a graph attention module with brightness temperature constraints to extract band features. Following this, the extracted band features undergo a comprehensive analysis through a multi-scale convolution module consisting of convolution kernels with multiple sizes, which has more variety and larger receptive fields for calculating the correlation between different bands features, assigning different weights to each band. Finally, a weight selection module is designed to filter the bands based on their assigned weights, creating a subset of bands with greater significance for LST retrieval. Training the designed model, 65100 observations are simulated utilizing MODTRAN, 80% allocated for training and 20% for testing. The experimental results validate the effectiveness of the proposed model, with a Root Mean Square Error (RMSE) of 1.85 K in practical applications on IASI imagery. This accomplishment substantiates the model’s capacity to reliably employ a judiciously selected subset of thermal infrared hyperspectral bands for LST retrieval applications, thus offering a promising contribution to the advancement of thermal infrared hyperspectral image processing methodologies.

Hyperspectral images provide spectral curves that reflect the "fingerprint" properties of substances, making them suitable for many applications. Thanks to the rapid development of computing resources, deep learning algorithms can significantly improve the cognitive ability of the network by extracting hidden features, and have been successfully applied to hyperspectral image processing, such as classification and detection. In this paper, a new hyperspectral target detection model based on one-dimensional generative adversarial networks (1D-GAN) is proposed. The proposed 1D-GAN network is designed to extract HSI features, and the probability is calculated accordingly whether the pixel to be detected is a target or background. In order to capture the spatial features, the guided filter is then used to obtain the final detection map. Performance comparison with several state-of-the-art methods demonstrate the effectiveness and efficiency of the proposed 1D-GAN algorithm.

Hyperspectral image (HSI) has attracted much attention because of its rich spectral information. However, due to the limitation of imaging hardware conditions, it is often difficult to directly obtain a high spatial resolution hyperspectral image (HR-HSI). To improve the resolution, it is an economical and effective method to fuse the hyperspectral image with the high spatial resolution multispectral image (HR-MSI) collected from the same scene. In recent years, with the development of deep learning, the convolutional neural network (CNN) based models have been applied to solve the super-resolution reconstruction of hyperspectral images. However, limited by the convolution kernel size, the receptive field of CNN is relatively small with more attention to the local information of the image. In order to solve this problem, this paper proposes a Swin Transformer based super-resolution reconstruction (STSR) network for hyperspectral images. Specifically, Swin Transformer structure is innovatively used in STSR as the skeleton of the network, where the Swin Transformer residuals are used to extract the global spatial feature information in the image. In addition, in order to retain the spectral details in the process of super-resolution reconstruction, a spectral attention module is introduced to preserve the original spectral information. The experimental results show that the high-resolution hyperspectral images fused by the proposed STSR method are superior to the comparison method in terms of vision and quality, which proves the superiority of this method.

Multi-exposure image fusion (MEF) technique is the most widely used method to obtain high dynamic range (HDR) images. Inspired by the recent successful application of Transformer in image processing, a hybrid dense connection network based on CNN and Transformer is proposed for MEF in this paper. Considering the importance of texture details to the multi-exposure image fusion task, shallow features containing rich texture details is also added to each dense layer, which are extracted by the pre-trained RepVGG. In addition, the dynamic weight calculation module is improved, so that different source images can obtain finer weight in the calculation of the loss function. Experiments are conducted on the dataset provided by MEFB, and both qualitative and quantitative comparisons show that the proposed method can achieve better results compared with the state-of-the-art algorithms.

Land surface temperature (LST) plays a crucial role in the physical and chemical processes of the land–atmosphere system. Remote sensing technology has greatly advanced the measurement of thermal infrared LST (TIR LST), which is the most widely utilized surface temperature product. However, cloud cover and mist often cause significant data loss in TIR LST. To address this issue and reconstruct the MYD11A1 LST under cloudy conditions, this study proposes an all-weather LST generation method based on the extreme gradient boosting (XGBoost) model. This method incorporates spatial-seamless passive microwave LST (PMW LST) to capture the nonlinear relationship between TIR LST and other variables. Compared to the MYD11A1 LST, the generated all-weather LST provides continuous spatial texture information without a significant boundary reconstruction effect, improving the accuracy of spatiotemporal variations in LST in China. In situ validation demonstrated the high accuracy of the generated all-weather LST, with mean$R^{2}$, bias, and unbiased root-mean-square error (ubRMSE) of 0.96 (0.91), 1.08 K (3.61 K), and 2.92 K (4.54 K) under clear (cloudy) daytime conditions, and 0.92 (0.95), −0.93 K (−2.96 K), and 3.09 K (3.04 K) under clear (cloudy) nighttime conditions. These results indicate the feasibility and reasonableness of the all-weather LST generation method developed in this study and affirm its ability to generate highly accurate all-weather LST.

Deep learning-based hyperspectral target detection (HTD) methods are limited by the lack of prior information. Self-supervised learning is a kind of unsupervised learning, which mainly mines its own self-supervised information from unlabeled data. By training the model with such constructed valid posterior information, a valuable representation model can be learned and can get rid of the dependence of deep models on prior information. To this end, this article proposes a self-supervised spectral-level contrastive learning-based HTD (SCLHTD) method to train a model with spectral difference discrimination capability for HTD in a self-supervised manner. First, the hyperspectral images (HSIs) to be detected are sampled in odd and even bands, and the obtained band subsets are then used to train the corresponding adversarial convolutional autoencoders. Feature extraction part of the trained encoder is then used as the data augmentation function, where the positive and negative pairs are constructed through data augmentation, and the backbone is used to extract the representative vectors of the augmented samples. Second, the representative vectors are mapped to the spectral contrast space using spectral contrastive head, where the similarity and dissimilarity of spectra are learned by maximizing the similarity of positive pairs while minimizing the similarity of negative pairs, so that the backbone can discriminate spectral differences. Finally, aiming at suppressing the background, edge-preserving filters are used in conjunction with space information to process the detection results acquired by utilizing spectrum information via cosine similarity to generate the final detection results. Experimental results illustrate that the proposed SCLHTD method can achieve superior performances for HTD.

A novel multi-scale fusion maximum entropy subspace clustering (MFMESC) for hyperspectral image (HSI) band selection is proposed in this paper. Subspace clustering is combined as a self-expression layer with stacked convolutional autoencoder, so that subspace clustering working in linear subspaces can deal with complicated HSI data with nonlinear characteristics. Multiple fully-connected linear layers are inserted between the encoder layers and their corresponding decoder layers to promote learning more favorable representations for subspace clustering. A multi-scale fusion module is designed to guide the fusion of multi-scale information extracted from different layers to learn a more discriminative self-expression coefficient matrix. Furthermore, the maximum entropy regularization is introduced in the subspace clustering to promote the connectivity within each subspace. Experimental results demonstrate the superiority of the proposed model against state of-the-art methods.

A new two-step physical method was developed to retrieve the land surface temperature (LST) from infrared atmospheric sounding interferometer (IASI) observations. This method relinearized the radiative transfer equation (RTE) by the tangents around the initial estimates of the LST, land surface emissivity (LSE), atmospheric equivalent temperature ($Ta$), and water vapor content ($q$). The Tikhonov regularization method and discrepancy principle (DP) iteration algorithm were employed to stabilize the ill-posed problem and obtain the final maximum likelihood solution of the LST with updating the initial estimation of LST, LSE,$Ta$, and$q$. A new channel selection scheme was proposed for this physical method to obtain an accurate LST estimation. This physical-based algorithm was tested on both simulated and real data obtained from the IASI. The root-mean-square error (RMSE) of the simulated LST is ~1 K based on an initial LST estimate with an RMSE of 2 K (1.9 K). The sensitivity analysis shows that the LST retrieval accuracy is ~1 K based on an LST with a random error of 3 K, constant initial LSE (0.97), 10%$Ta$error, and 40%$q$error. Compared with the Advanced Very High Resolution Radiometer onboard Metop (AVHRR/Metop) LST product, the physical method achieves the LST retrieval accuracy of 1.5 and 1 K for real daytime and nighttime IASI data obtained in the study area. Based on the new method, the LST can be retrieved with an accuracy similar to that of the AVHRR/Metop LST product.

Despite continuing to progress in hyperspectral image classification (HSIC) based on deep learning, the classification accuracy is limited to furtherly improve in the absence of labeled samples. To address this issue, the metric-based prototypical networks for few-shot learning have enjoyed widespread popularity. However, the conventional prototypical networks are vulnerable to the selected examples and fail to accomplish representative predictions for the prototypes in complicated situations. In this paper, we propose a multi-view calibrated prototype-learning framework for few-shot HSIC, which consists of three rectified strategies from different views to improve the robustness of prototypes in the embedding space. Specifically, the calibrated aggregation network is the first presented to calibrate the representations with local patches aggregation for the enhancement of the prototypes. Moreover, to improve the compactness of the intraclass expression, the calibrated metric learning with regularization terms is designed to strengthen the discrimination of the prototypes. Furthermore, we calibrate the feature distribution of supervised samples by transferring statistical knowledge to eliminate the local bias in the test phase. The extensive experimental results and analysis of three hyperspectral image datasets demonstrate the superiority of the proposed architecture compared with other advanced methods.

Hyperspectral image (HSI) provided with a substantial number of correlated bands causes calculation consumption and an undesirable "dimension disaster" problem for the classification. Band selection (BS) is an effective measure to reduce the information redundancy with the physics spectrum preserved for HSI. Although the existing BS methods have achieved noticeable progress, the correlation between neighbor bands still needs to be mined deeply for an effective selection criterion. This paper proposes a BS approach to collecting the discriminative band subset for hyperspectral image classification (HSIC), which adopts the self-supervised learning paradigm to implement the BS by auxiliary spectrum rebuilding task. In specific, we utilized a Convolutional neural network (CNN) and Graph Convolutional Network (GCN) for the spectral-spatial feature extraction. Next, GCN and CNN are developed for the refinement of the band correlation sequentially. Afterward, the selected bands in terms of the acquired correlation are fed into the presented self-supervised spectrum rebuilding network for spectral reconstruction. Simultaneously, the proposed architecture completed the selection with the optimization of the band reconstruction by a defined loss function. In this way, we supply substitution for selection criterion and path searching through the end-to-end framework. The extensive experimental results and analysis demonstrated that the proposed hybrid architecture provided a competitive band subset for the classification, and the accuracies with different types of classifiers are more effective than the compared BS methods.

High dynamic range (HDR) technology is one of the most widely used ways to improve image quality, and fusion of a series of low dynamic range (LDR) images is the main measure to obtain a HDR image. However, because moving objects are often found in a series of LDR images, the fused HDR images produce ghostly shapes. In order to eliminate ghosts, this paper proposes a ghost-free multi-exposure fusion method. Firstly, aligning the moving object in the input multi-exposure sequence images with the moving object in the reference image, and the aligned sequence images are obtained. In order to consider assigning more weight to pixels in the better exposure area, two weighting functions are defined. One is to measure pixel values relative to the overall brightness and adjacent exposure images, and the other one is to reflect pixel values within a range that has a larger global gradient relative to other exposures. Based on these two weighting functions, the low exposure sequence images aligned in the Laplacian pyramid are finally fused. Through experimental comparison, the obtained image has no ghost, good visual effect, and rich details.

This paper develops a new approach to band selection for multiple-target detection, called constrained-target band selection (CTBS). Its idea is derived from the concept of constrained energy minimization (CEM) by constraining a target of interest, while minimizing the variance resulting from the background (BKG). By taking advantage of CEM, the variance produced by a target of interest can be further used as a measure of prioritizing bands as well as a means of selecting bands for this particular target. As a result, two CTBS-based band prioritization (BP) criteria, called minimal variance-based BP (MinV-BP) and maximal variance-based BP (MaxV-BP), and two CTBS-based BS methods, called sequential forward CTBS (SF-CTBS) and sequential backward CTBS (SB-CTBS), can be derived for multiple-target detection. Since the bands selected by CTBS vary with targets of interest used to constrain CEM, in order for CTBS to be applied to multiple targets, a new fusion technique, called band fusion selection (BFS), is further developed for CTBS to integrate bands selected by different targets so that CTBS can work for all targets. Unlike most BS methods for target detection which generally simultaneously select a fixed set of bands for all targets of interest, the ideas of constraining multiple-target detection and using BFS are novelty of this paper. Experimental results show that CTBS performs well for multiple-target detection.