Hyperspectral image feature extraction plays a crucial role in reducing the redundancy and correlation among spectral bands while preserving the essential information. Knowledge-driven feature extraction methods, such as spectral indices (SIs), leverage the interaction mechanisms between electromagnetic waves and materials to enhance the characteristic attributes of ground objects through band operations. These methods offer key advantages, including strong physical interpretability, simple construction, and robust scene reusability. However, most of the existing SIs still rely on expert knowledge tailored to specific scenarios, leading to inherent limitations, such as subjectivity, high time consumption, and implementation complexity. In this article, to address these challenges, we propose a hyperspectral image multi-agent deep reinforcement learning feature extraction (HI-MAFE) algorithm, aiming to alleviate the burden of manual SIs design by human experts. HI-MAFE employs a heuristic “generation-selection” strategy to simulate the decision-making process of domain experts, with specifically designed deep reinforcement learning (DRL) models for both the generation and selection steps. To accelerate exploration in a high-dimensional action space, the model incorporates a multi-agent deep reinforcement learning (MADRL) framework. The experimental results demonstrate the effectiveness and superiority of the proposed algorithm for hyperspectral image classification. The proposed HI-MAFE framework leverages DRL to autonomously generate meaningful environmental interpretation from spectral data, thereby reducing the reliance on manually designed SIs. This research can inspire future work in SIs construction and complement the limitations of data-driven approaches.

Soil organic matter (SOM) content is a crucial indicator for assessing soil fertility and serves as a key factor in sustaining a viable agricultural system. With the continuous advancement and improvement of remote sensing technology, hyperspectral imagery has been employed for the monitoring of SOM. While the numerous bands reveal finer details within the spectral features, this also brings information redundancy and noise interference. Currently, the dimensionality reduction methods designed for hyperspectral imagery encounter difficulties in achieving optimal band combinations. As a result, swiftly and accurately capturing the spectral features of SOM becomes a challenging task. In this article, aiming to address the inefficiency and instability in hyperspectral feature selection, we propose a metaheuristic-based algorithm—the improved weighted marine predators algorithm (IWMPA)—for hyperspectral feature selection. Specifically, we simulated the process of hyperspectral feature selection using the foraging strategy of marine predators. We employed prior weight coefficients and reverse learning operations in the initialization phase to accelerate the convergence of the population and introduced mutation operations into the phase of development to prevent the occurrence of local optima traps. We employed the IWMPA feature selection method to establish SOM estimation models within the research area of Yitong Manchu Autonomous County in China. The results demonstrated that the hyperspectral features selected using the IWMPA approach yield favorable outcomes in the SOM estimation models. Specifically, in the best-performing regression model of this study, R2 on the test set was 0.7225. These experimental results suggest that, in comparison to the existing methods, the proposed IWMPA method is more adept at capturing the spectral features of SOM.

Automatic building localization and extraction based on high-resolution remote sensing images is of great importance to city mapping and smart city management. Extracted buildings with high precisions and fine boundaries contribute to the vectorization operation and, thus, the digital line graph (DLG) production. In this regard, a comprehensive framework Poly-BRBLE is proposed, combining a boundary refinement-based individual building localization and extraction model BRBLE as well as a particularly revised regularization method. The BRBLE is mainly composed of a multiscale feature fusion and propagation module and a coarse-to-fine mask optimization module, which allow the model to identify buildings from similar backgrounds and extract them with precise boundaries. Comparisons were made between BRBLE and other classical and state-of-the-art models on the WHU building dataset, the Chinese building instance segmentation dataset, and the Inria Polygon dataset, which demonstrated that BRBLE outperformed the second-best model by 2%, 1.5%, and 1.2%, respectively, in${\text {AP}}_{\text {mask}}$, and the advantage was further enlarged on large objects to 5.2%, 1.8%, and 2%. Moreover, we built the SH building dataset, which contained complex-shaped buildings and mixed-built environments, where it was demonstrated that BRBLE outperformed the second-best model by 3.4% in${\text {AP}}_{\text {mask}}$. We further compared the precisions of the building footprints obtained by the Poly-BRBLE and other different methods, where Poly-BRBLE showed a superior performance, with an${\text {AP}}_{\text {mask}}$score of 60.4%, which demonstrated that it is capable of extracting complicated buildings, such as high-rise buildings and villas, and factories of multiple shapes, even though the images were off-nadir.

Hyperspectral target detection (HTD) relies on prior target spectra to locate the targets of interest within hyperspectral images. Recently, deep learning methods have shown their potential in hyperspectral feature extraction and multi-scale feature fusion. In this article, we propose a background-aware sparse transformer network (BASTNet) for hyperspectral target detection, to solve the problems of target sample imbalance and underutilization of global information. Firstly, the proposed method utilizes random masking and target spectra generation strategies to establish an image-level training paradigm, constructing sufficient and balanced training samples to prompt the network to learn spatial-contextual features between the target and the background. We then introduce a Siamese sparse transformer network (S2TNet) with an encoder-decoder structure to achieve fast inference for large-scene hyperspectral imagery. Specifically, S2TNet consists of pyramid feature extraction, multi-scale feature fusion, and a target detector, with a sparse self-attention mechanism enhancing the focus on target regions and improving the separability between target and background. Furthermore, a background-aware learning mechanism is introduced that uses a foreground and background guidance loss that attenuates the interference of background noise on the target detection. Experiments on five benchmark datasets demonstrate the superiority and applicability of the proposed BASTNet method, showing that it outperforms the current state-of-the-art hyperspectral target detection methods.

Semi-supervised change detection(SSCD) in remote sensing faces two critical challenges: cyclical error propagation from noisy pseudo-labels and the loss of fine-grained boundary details. To address these, we propose a novel SSCD framework, the Semi-supervised Connectivity-Aware Decoupling Network (SCAD-Net). SCAD-Net breaks the cycle of error amplification with a strategy called Pixels-Regions via Curriculum-guided Consistency Learning (P2R-CL). This strategy progresses from initial, high-confidence pixel-level supervision to more flexible, region-based semantic correction as training matures. To tackle boundary ambiguity, SCAD-Net employs a two-stage decoupling architecture: a Channel Information Decoupling Module (CIDM) separates categorical and directional features, followed by a Multi-scale Attention and Feature Integration (MAFI) module that reintegrates this information for precise boundary localization. Experiments on four benchmark datasets demonstrate our method’s superiority. On the SHD, using only 10% labeled data, SCAD-Net achieves an F1-score of 90.1%, representing a 3.88% improvement over state-of-the-art methods. Similarly, strong performance is observed on the LEVIR-CD (90.31% F1), WHU-CD (89.35% F1) and CDD (87.05% F1) datasets. This feature decoupling paradigm offers a robust approach for semi-supervised remote sensing in environments.

Convolutional neural networks (CNNs) have achieved success in HSI classification, but the performance is constrained by the limited reception field. In this regard, vision transformer is introduced recently, which is of powerful capabilities in long-range feature extraction for HSI classification. However, transformers are computation intensive and poor for local feature extraction. The motivation for this study is to build a lightweight hybrid model, which ensembles the respective inductive bias from CNNs and global receptive field from transformers. In this work, we propose a concise and efficient framework—the spatial-spectral feature extraction network with patch attention module (PASSNet), to simultaneously extract both local and global features. Specifically, we design an innovative plugin called patch attention module (PAM), which can be easily integrated into both CNNs and transformers blocks to extract spatial-spectral features from multiple spatial perspectives. Besides, a novel partial convolution operation is introduced, with a reduced computational cost than vanilla convolution operation. Through coupling the local attention from the CNNs with the global receptive fields in the transformers, the proposed PASSNet exhibits a superior classification performance on three well-known datasets with a small training sample size.

The issues of spectral redundancy and limited training samples hinder the widespread application and development of hyperspectral images. In this letter, a novel active deep feature extraction scheme is proposed by incorporating both representative and informative measurement. Firstly, an adversarial autoencoder is modified to suit the classification task with deep feature extraction. Dictionary learning and a multi-variance and distributional distance (MVDD) measure are then introduced to choose the most valuable candidate training samples, where we use the limited labeled samples to obtain a high classification accuracy. Comparative experiments with the proposed querying strategy were carried out with two hyperspectral datasets. The experimental results obtained with the two datasets demonstrate that the proposed scheme is superior to the others. With this method, the unstable increase in accuracy is eliminated by incorporating both informative and representative measurement.

The ZY1-02D satellite, which was launched in 2019, is China’s first civil hyperspectral satellite. However, the laboratory calibration and vicarious calibration methods could not provide accurate radiometric calibration coefficients after the satellite had been launched. In this article, we describe how a cross-calibration method was utilized to calibrate the ZY1-02D hyperspectral imager using the well-calibrated Gaofen-5 Advanced Hyperspectral Imager (GF-5 AHSI). The 6S radiative transfer model was selected to simulate the apparent reflectance of the two hyperspectral sensors under corresponding imaging conditions, and the calibration coefficients were calculated by spectral channel matching. The reflectance-based vicarious calibration was carried out for comparison. Through the validation experiments, it is shown that the reflectance data obtained by cross-calibration and vicarious calibration are basically consistent, showing a stable radiation performance. At the Dunhuang calibration site, the ratio of measured surface reflectance to the cross-calibrated image reflectance is between 0.9 and 1.1, the$R^{2}$values are more than 0.96, and the spectral angles are less than 3°. The validation results for different ground features also show the applicability of the corrected coefficients. When compared with different sensors, the maximum difference between the ZY1-02D reflectance results after cross-calibration and Landsat-8/Sentinel-2 is less than 0.04 and the mean difference is less than 0.02, which further proves that the ZY1-02D hyperspectral imager has a high radiation accuracy after cross-calibration. The proposed cross-calibration method could be used as an effective supplement to the on-orbit calibration method and could also be extended to other satellite hyperspectral imagers.

The highly correlated spectral features and the limited training samples pose challenges in hyperspectral image classification. In this article, to tackle the issues of end-to-end feature learning and transfer learning with limited labeled samples, we propose a unified multiscale learning (UML) framework, which is based on a fully convolutional network. A multiscale spatial-channel attention mechanism and a multiscale shuffle block are proposed in the UML framework to improve the problem of land-cover map distortion. The contextual information and the spectral feature are enhanced before the last classification layer based on three strategies in this work: 1) the channel shuffle operation, which was employed to learn the more effective spectral characteristics by disordering the channels of the feature map; 2) multiscale block, which considered the contextual information in multiple ranges; and 3) spatiospectral attention, which enhanced the expression of the important characteristic among all pixels. Three hyperspectral datasets, including two airborne hyperspectral images and one spaceborne hyperspectral image, were used to demonstrate the performance of the UML framework in both classification and transfer learning. The experimental results confirmed that the proposed method outperforms most of the state-of-the-art hyperspectral image classification methods. The source code is released athttps://github.com/Hyper-NN/UML.

The visible-shortwave infrared Advanced Hyperspectral Imager (AHSI) is a payload onboard the Gaofen-5 satellite, which is China’s first hyperspectral satellite and is part of the Chinese High-Resolution Earth Observation System. As a supplement to the onboard radiometric calibration of the AHSI instrument, vicarious calibration is also required, which is independent of the instrument-based calibration. In this article, a reflectance-based vicarious calibration approach is presented, which takes surface reflectance data, aerosol data, and atmospheric water vapor data into account. The Dunhuang test site, which is one of the China Radiometric Calibration Sites (CRCS) for the vicarious calibration of spaceborne sensors, possesses stable, uniform, and measurable surface objects, so it was chosen as the radiation source to replace the laboratory and onboard calibrators. A Spectra Vista Corporation (SVC) spectral radiometer and a CE318 sun photometer were utilized for the measurement of the surface reflectance and the condition of the aerosol, respectively. The radiance at the entrance pupil at the top of atmosphere was then obtained through the MODerate resolution atmospheric TRANsmission (MODTRAN) atmospheric transmission model. The surface reflectance was obtained using the Fast Line-of-sight Atmospheric Analysis of Hypercubes (FLAASH) atmospheric model for validation. The results show that, with regard to the calibration coefficients, the calibrated AHSI instrument presents a stable radiometric performance among different land-cover types. The ratios on all the bands are between 0.8 and 1.2 and are consistent with the reflectance data from the Dunhuang test site. The${R} ^{{2}}$values are all greater than 0.95 and the spectral angle is all less than 2°. The standard deviations of the ratios are less than 3% for each chosen band, which proves that the calibrated data have a high consistency with thein situmeasurements. When compared with Landsat 8 and Sentinel-2, the mean errors of the surface reflectance are all under 0.06, which further demonstrates that the calibrated reflectance has a high accuracy.

Heavy metals affect the soil and food adversely, as well as threaten the human health and ecosystem by biologically cumulative effect. Yitong county, an important grain production base with abundance of black soil, has been threatened by a gold mine in fields of the security of food production and environment. For the exploration of the pollution situation, the heavy metal contents in agricultural soil around the gold-mining area have been determined in this work. Firstly, a total of 91 samples were collected by field surveying, and the concentrations of 7 elements (As, Cd, Cr, Cu, Ni, Pb and Zn) in those soil samples were analyzed. After that, geoaccumulation index (Igeo), pollution index (PI) and potential ecological risk (PER) were selected as the evaluation indexes to assess heavy metal pollution. In addition, the varimax normalized rotation was applied to trace the source of pollution. The results indicate that the study area was contaminated on a moderate level.

Deep generative models such as the generative adversarial network (GAN) and the variational autoencoder (VAE) have obtained increasing attention in a wide variety of applications. Nevertheless, the existing methods cannot fully consider the inherent features of the spectral information, which leads to the applications being of low practical performance. In this article, in order to better handle this problem, a novel generative model named the conditional variational autoencoder with an adversarial training process (CVA2E) is proposed for hyperspectral imagery classification by combining variational inference and an adversarial training process in the spectral sample generation. Moreover, two penalty terms are added to promote the diversity and optimize the spectral shape features of the generated samples. The performance on three different real hyperspectral data sets confirms the superiority of the proposed method.

The increase in the spectral and spatial information of hyperspectral imagery poses challenges in classification due to the fact that spectral bands are highly correlated, training samples may be limited, and high resolution may increase intraclass difference and interclass similarity. In this paper, in order to better handle these problems, a Caps-TripleGAN framework is proposed by exploring the 1-D structure triple generative adversarial network (TripleGAN) for sample generation and integrating CapsNet for hyperspectral image classification. Moreover, spatial information is utilized to verify the learning capacity and discriminative ability of the Caps-TripleGAN framework. The experimental results obtained with three real hyperspectral data sets confirm that the proposed method outperforms most of the state-of-the-art methods.

With the Geographic Information System (GIS) analytical tools and Remote Sensing technology, the main territory's Net Primary Productivity (NPP) in China from 2001 to 2010 is estimated based on CASA model. The analysis of the temporal and spatial variations in NPP shown that NPP in the southern regions shows increase trend after the first five years' decrease, in addition, the southern region of China generally has higher NPP as the northwest is significantly lower than other regions Moreover, the influence of climatic changes on the NPP's evolution represent that the NPP have different influent factors in different regions and periods.