Spatial Spectral Information Research Articles

A joint model based on graph and deep learning for hyperspectral anomaly detection

Through the years, graph theory has gradually been applied in hyperspectral image (HSI) processing. The graph theory method does not need to consider the structural characteristics of the dataset and has achieved satisfactory application results. However, the current graph theory methods applied in HSI processing mainly focus on spectral characteristics. This paper proposes a joint model based on graph and deep learning (JGD) for hyperspectral anomaly detection (AD). The proposed JGD consists of a spectral sub-model using graph and a spatial sub-model using deep learning. In the spectral sub-model, graph Fourier transform (GFT) is used, and the reconstruction errors between the original HSI and the HSI after GFT are mapped to fractional Fourier domain (FrFD) for AD results, which mainly makes use of the global spectral information. In the spatial sub-model, stacked autoencoders (SAEs) and an adaptive algorithm are combined to obtain AD results, which mainly use local spatial information and spectral information. These two sub-models are combined through the differential fusion method to form the JGD, which fully utilizes spectral and spatial information of hyperspectral images (HSIs). Compared with seven other algorithms on four real HSIs, the proposed JGD shows superior AD performance.

PDRs4All

Context. JWST has taken the sharpest and most sensitive infrared (IR) spectral imaging observations ever of the Orion Bar photodis-sociation region (PDR), which is part of the nearest massive star-forming region the Orion Nebula, and often considered to be the ‘prototypical’ strongly illuminated PDR. Aims. We investigate the impact of radiative feedback from massive stars on their natal cloud and focus on the transition from the H II region to the atomic PDR – crossing the ionisation front (IF) –, and the subsequent transition to the molecular PDR – crossing the dissociation front (DF). Given the prevalence of PDRs in the interstellar medium and their dominant contribution to IR radiation, understanding the response of the PDR gas to far-ultraviolet (FUV) photons and the associated physical and chemical processes is fundamental to our understanding of star and planet formation and for the interpretation of any unresolved PDR as seen by JWST. Methods. We used high-resolution near-IR integral field spectroscopic data from NIRSpec on JWST to observe the Orion Bar PDR as part of the PDRs4All JWST Early Release Science programme. We constructed a 3″ × 25″’ spatio-spectral mosaic covering 0.97– 5.27 μm at a spectral resolution R of ~2700 and an angular resolution of 0.075″–0.173″. To study the properties of key regions captured in this mosaic, we extracted five template spectra in apertures centred on the three H2 dissociation fronts, the atomic PDR, and the H II region. This wealth of detailed spatial-spectral information was analysed in terms of variations in the physical conditions-incident UV field, density, and temperature – of the PDR gas. Results. The NIRSpec data reveal a forest of lines including, but not limited to, He I, H I , and C I recombination lines; ionic lines (e.g. Fe III and Fe II); O I and N I fluorescence lines; aromatic infrared bands (AIBs, including aromatic CH, aliphatic CH, and their CD counterparts); pure rotational and ro-vibrational lines from H2; and ro-vibrational lines from HD, CO, and CH+, with most of them having been detected for the first time towards a PDR. Their spatial distribution resolves the H and He ionisation structure in the Huygens region, gives insight into the geometry of the Bar, and confirms the large-scale stratification of PDRs. In addition, we observed numerous smaller-scale structures whose typical size decreases with distance from θ1 Ori C and IR lines from C I , if solely arising from radiative recombination and cascade, reveal very high gas temperatures (a few 1000 K) consistent with the hot irradiated surface of small-scale dense clumps inside the PDR. The morphology of the Bar, in particular that of the H2 lines, reveals multiple prominent filaments that exhibit different characteristics. This leaves the impression of a ‘terraced’ transition from the predominantly atomic surface region to the CO-rich molecular zone deeper in. We attribute the different characteristics of the H2 filaments to their varying depth into the PDR and, in some cases, not reaching the C+/C/CO transition. These observations thus reveal what local conditions are required to drive the physical and chemical processes needed to explain the different characteristics of the DFs and the photochemical evolution of the AIB carriers. Conclusions. This study showcases the discovery space created by JWST to further our understanding of the impact radiation from young stars has on their natal molecular cloud and proto-planetary disk, which touches on star and planet formation as well as galaxy evolution.

MEA-EFFormer: Multiscale Efficient Attention with Enhanced Feature Transformer for Hyperspectral Image Classification

Hyperspectral image classification (HSIC) has garnered increasing attention among researchers. While classical networks like convolution neural networks (CNNs) have achieved satisfactory results with the advent of deep learning, they are confined to processing local information. Vision transformers, despite being effective at establishing long-distance dependencies, face challenges in extracting high-representation features for high-dimensional images. In this paper, we present the multiscale efficient attention with enhanced feature transformer (MEA-EFFormer), which is designed for the efficient extraction of spectral–spatial features, leading to effective classification. MEA-EFFormer employs a multiscale efficient attention feature extraction module to initially extract 3D convolution features and applies effective channel attention to refine spectral information. Following this, 2D convolution features are extracted and integrated with local binary pattern (LBP) spatial information to augment their representation. Then, the processed features are fed into a spectral–spatial enhancement attention (SSEA) module that facilitates interactive enhancement of spectral–spatial information across the three dimensions. Finally, these features undergo classification through a transformer encoder. We evaluate MEA-EFFormer against several state-of-the-art methods on three datasets and demonstrate its outstanding HSIC performance.

A fast hypergraph neural network with detail preservation for hyperspectral image classification

ABSTRACT Hypergraph neural networks (HGNNs), extending the techniques of graph neural networks, have been applied to various fields due to their ability to capture more complex high-order node relationships. However, for hyperspectral image (HSI) classification tasks, previous HGNN-based works usually constructed hypergraphs using pixels as nodes, resulting in massive computational costs. Meanwhile, pixel-level personalized features are required for HSI classification. To achieve high efficiency and accuracy simultaneously, this paper presents a fast hypergraph neural network with detail preservation (DPFHNet) for HSI classification. It constructs hypergraphs at the superpixel level to reduce time consumption and supplement pixel-level detail features through a classification refinement module. This framework contains multiple stages. Firstly, its main stage is designed with HGNNs from a superpixel viewpoint rather than pixels, providing a fast strategy to capture high-order complex relationships of multiple homogeneous irregular regions. After that, auxiliary stages based on convolutional neural networks are integrated into the main stage, which adopts a hierarchical design and attempts to acquire pixel-level spatial-spectral information before the hypergraph feature extraction of the main stage, assisting in learning more valuable features. Finally, a classification refinement module is constructed, which generates pixel-level detail features to refine the superpixel-level features obtained by HGNN. Experiments on three datasets illustrate that DPFHNet achieves competitive results and efficiency compared to advanced methodologies.

An improved arithmetic optimization algorithm with multi-strategy for adaptive multi-spectral image fusion

Panchromatic and multi-spectral image fusion, called panchromatic sharpening, is the process of combining the spatial and spectral information of the source image into the fused image to give the image a higher spatial and spectral resolution. In order to improve the spatial resolution and spectral information quality of the image, an adaptive multi-spectral image fusion method based on an improved arithmetic optimization algorithm is proposed. This paper proposed improved arithmetic optimization algorithm, which uses dynamic stochastic search technique and oppositional learning operator, to perform local search and behavioral complementation of population individuals, and to improve the ability of population individuals to jump out of the local optimum. The method combines adaptive methods to calculate the weights of linear combinations of panchromatic and multi-spectral gradients to improve the quality of fused images. This study not only improves the quality and effect of image fusion, but also focuses on optimizing the operation efficiency of the algorithm to have real-time and high efficiency. Experimental results show that the proposed method exhibits strong performance on different datasets, improves the spatial resolution and spectral information quality of the fused images, and has good adaptability and robustness. The source code is available at: https://github.com/starboot/IAOA-For-Image-Fusion.

Prediction of species richness and diversity in sub‐alpine grasslands using satellite remote sensing and random forest machine‐learning algorithm

AbstractAimsRemote‐sensing approaches could be beneficial for monitoring and compiling essential biodiversity data because they are cost‐effective and allow for coverage of large areas over a short period. This study investigated the relationship between multispectral remote‐sensing data from Landsat 8 and Sentinel‐2 and species richness and diversity in mountainous and protected grasslands.LocationsGolden Gate Highlands National Park, Free State, South Africa.MethodsIn‐situ data of plant species composition and cover from 142 plots with 16 relevés each were distributed across the study site and used to calculate species richness and Shannon–Wiener species diversity index (species diversity). We used a machine‐learning random forest algorithm to optimize the prediction of species richness and diversity. The algorithm was used to identify the optimal spectral bands and vegetation indices for estimating species richness and diversity. Subsequently, the selected bands and vegetation indices were used to estimate species richness through random forest regression.ResultsThis research found weak relationships between remote‐sensing vegetation indices and the diversity metrics, but significant relationships were found between some spectral bands and diversity metrics. Moreover, using machine‐learning random forest, the multispectral data sets exhibited strong predictive powers. In this investigation, near‐infrared (NIR) seemed to be the most selected band for both sensors to explain species diversity in mountainous grasslands.Main conclusionsThis finding further ascertains the efficiency of optimizing high spatial resolution spectral information to estimate plant species richness and diversity. This research shows that NIR, Soil‐Adjusted Vegetation Index (SAVI) and Enhanced Vegetation Index (EVI) are the most adequate for predicting species richness and diversity in mountainous grasslands with relatively good accuracies. Plant phenology and the choice of sensor affect the relationship between spectral information and species diversity variables.

3D-GhostNet: A novel spatial-spectral algorithm to improve foodborne bacteria classification coupled with hyperspectral microscopic imaging technology

A novel 3D-GhostNet has been developed for hyperspectral microscopic imaging (HMI) data analysis to improve classification of foodborne pathogens. The HMI, leveraging microscopic technology, surpasses traditional spectral imaging in terms of sensitivity and resolution by utilizing visible/near-infrared spectroscopy integrated with high-resolution single-cell images. The newly constructed 3D-Ghost network directly processes and recognizes single-cell hypercube, enables HMI technology to become near-real-time detection by combining with automated data processing pipeline. 3D-GhostNet incorporates Ghost modules as its backbone and adds convolutional block attention module (CBAM) for adaptively extracting high-dimensional depth spatial-spectral information. In the task of identifying four different foodborne bacteria cells, our 3D-GhostNet achieved 100% accuracy, surpassing spectral-only-based (SOB) classifiers such as linear discriminant analysis (LDA, 89.7%), support vector machine (SVM, 93.9%), 1D convolutional neural network (1D-CNN, 98%), and long short-term memory (LSTM, 98.5%). Furthermore, our independent blind test results indicate that 3D-GhostNet remains robust for identifying complex mixtures of pathogens. As a label-free detection tool, 3D-GhostNet-assisted HMI technology is fully competent in classification of different foodborne pathogens, suggesting broad applications in food safety and quality.

A hybrid convolution transformer for hyperspectral image classification

ABSTRACT Hyperspectral images play a crucial role in remote sensing applications surveillance, environment and precision agriculture, containing abundant object information. However, they often face challenges such as limited labelled data and imbalanced classes. In recent years, convolutional neural networks (CNNs) have shown impressive performance in computer vision tasks, including hyperspectral image classification. The emergence of transformers has also attracted attention for hyperspectral image analysis due to their promising capabilities. Nevertheless, transformers typically demand a substantial amount of training data, making their application challenging in scenarios with limited labelled samples. To overcome this limitation, we propose a hybrid convolution transformer framework. Our method uses a vision transformer and a residual 3D convolutional neural network model. It also uses a sequence aggregation layer to avoid overfitting issues that come up when there isn’t enough training data. Our proposed residual channel attention module captures richer spatial-spectral complementary information and maintains spectral details during the feature extraction process. We conducted experiments on three benchmark datasets. The proposed model achieved state of the art performance 99.75 % , 99.46 % and 99.95 % in terms of overall accuracy (OA) using only 5 % , 5 % and 10 % labelled training samples respectively. Which is better than other state of the art methods.

Exploring an application-oriented land-based hyperspectral target detection framework based on 3D–2D CNN and transfer learning

Target detection based on hyperspectral images refers to the integrated use of spatial information and spectral information to accomplish the task of localization and identification of targets. There are two main methods for hyperspectral target detection: supervised and unsupervised methods. Supervision method refers to the use of spectral differences between the target to be tested and the surrounding background to identify the target when the target spectrum is known. In ideal situations, supervised object detection algorithms perform better than unsupervised algorithms. However, the current supervised object detection algorithms mainly have two problems: firstly, the impact of uncertainty in the ground object spectrum, and secondly, the universality of the algorithm is poor. A hyperspectral target detection framework based on 3D–2D CNN and transfer learning was proposed to solve the problems of traditional supervised methods. This method first extracts multi-scale spectral information and then preprocesses hyperspectral images using multiple spectral similarity measures. This method not only extracts spectral features in advance, but also eliminates the influence of complex environments to a certain extent. The preprocessed feature maps are used as input for 3D–2D CNN to deeply learn the features of the target, and then, the softmax method is used to output and obtain the detection results. The framework draws on the ideas of integrated learning and transfer learning, solves the spectral uncertainty problem with the combined similarity measure and depth feature extraction network, and solves the problem of poor robustness of traditional algorithms by model migration and parameter sharing. The area under the ROC curve of the proposed method has been increased to over 0.99 in experiments on both publicly available remote sensing hyperspectral images and measured land-based hyperspectral images. The availability and stability of the proposed method have been demonstrated through experiments. A feasible approach has been provided for the development and application of specific target detection technology in hyperspectral images under different backgrounds in the future.

Multi-Scale Residual Spectral–Spatial Attention Combined with Improved Transformer for Hyperspectral Image Classification

Aiming to solve the problems of different spectral bands and spatial pixels contributing differently to hyperspectral image (HSI) classification, and sparse connectivity restricting the convolutional neural network to a globally dependent capture, we propose a HSI classification model combined with multi-scale residual spectral–spatial attention and an improved transformer in this paper. First, in order to efficiently highlight discriminative spectral–spatial information, we propose a multi-scale residual spectral–spatial feature extraction module that preserves the multi-scale information in a two-layer cascade structure, and the spectral–spatial features are refined by residual spectral–spatial attention for the feature-learning stage. In addition, to further capture the sequential spectral relationships, we combine the advantages of Cross-Attention and Re-Attention to alleviate computational burden and attention collapse issues, and propose the Cross-Re-Attention mechanism to achieve an improved transformer, which can efficiently alleviate the heavy memory footprint and huge computational burden of the model. The experimental results show that the overall accuracy of the proposed model in this paper can reach 98.71%, 99.33%, and 99.72% for Indiana Pines, Kennedy Space Center, and XuZhou datasets, respectively. The proposed method was verified to have high accuracy and effectiveness compared to the state-of-the-art models, which shows that the concept of the hybrid architecture opens a new window for HSI classification.

S[formula omitted]WaveNet: A novel spectral–spatial wave network for hyperspectral image classification

Deep learning has made significant progress in hyperspectral image (HSI) classification, and its powerful ability to automatically learn abstract features is well recognized. Recently, the simple architecture of multi-layer perceptron (MLP) has been extensively employed to extract long-range dependencies of HSI and achieved impressive results. However, existing MLP-based models exhibit insufficient representation of spectral–spatial information in HSI and generally aggregate features with fixed weights, which limits their ability to capture semantic differences. To tackle these challenges, this paper proposes a novel spectral–spatial wave network (S2WaveNet) for HSI classification tasks to enhance the representation capability of spectral–spatial features in ground objects. Specifically, the spectral–spatial wave mixer (S2WaveMixer) block is designed as a key component to represent each HSI input as a wave function with amplitude and phase parts. Thus, it enables a deeper dynamic perception and facilitates the extraction of spectral–spatial feature variations of ground objects. The amplitude represents the original features and the phase term is a complex value changing based on the semantic contents of the input images. Furthermore, the inception unit is introduced into the S2WaveMixer block to consider spectral–spatial information at multiple granularity levels. Experiments conducted on five public datasets demonstrate the superiority of S2WaveNet in classification performance and generalization compared to competitors.

Spectral-Spatial Feature Fusion for Hyperspectral Anomaly Detection.

Hyperspectral anomaly detection is used to recognize unusual patterns or anomalies in hyperspectral data. Currently, many spectral-spatial detection methods have been proposed with a cascaded manner; however, they often neglect the complementary characteristics between the spectral and spatial dimensions, which easily leads to yield high false alarm rate. To alleviate this issue, a spectral-spatial information fusion (SSIF) method is designed for hyperspectral anomaly detection. First, an isolation forest is exploited to obtain spectral anomaly map, in which the object-level feature is constructed with an entropy rate segmentation algorithm. Then, a local spatial saliency detection scheme is proposed to produce the spatial anomaly result. Finally, the spectral and spatial anomaly scores are integrated together followed by a domain transform recursive filtering to generate the final detection result. Experiments on five hyperspectral datasets covering ocean and airport scenes prove that the proposed SSIF produces superior detection results over other state-of-the-art detection techniques.

Multi-scale feature learning via residual dynamic graph convolutional network for hyperspectral image classification

ABSTRACT Thanks to the relevant spectral-spatial information, hyperspectral images (HSIs) have been widely exploited in Earth observation. Recently, graph convolutional networks (GCNs) have attracted increasing attention in HSI classification due to their advantages in processing non-Euclidean structure data. Unlike convolutional neural networks (CNNs), which perform convolution operations on regular square regions, GCNs can directly work on graph structure data to extract the relationships among adjacent land covers. However, extracting meaningful and deep discriminative spectral-spatial features from HSIs is still a challenging task. In this article, a novel multi-scale feature learning via residual dynamic graph convolutional network is designed for HSI classification, which can extract large-scale contextual spatial structures at superpixel-level graph and local spectral-spatial information at pixel-level, significantly improving the performance of HSI classification. Unlike from the existing GCN-based methods that operate on a graph with a fixed neighbourhood size, multiple graphs with diverse neighbourhood scales are built to comprehensively leverage spectral-spatial information and relationship at multiple scales, and these graphs are dynamically updated to generate more discriminative features (via dynamic GCN) during the convolution process. Moreover, to fully use the multi-scale features extracted from HSIs, a multi-scale feature fusion module is developed to emphasize important features and suppress irrelevant ones. Extensive experiments carried on three benchmark data sets demonstrate the superiority of the proposed approach over other state-of-the-art methods.

Hyperspectral Image Shadow Enhancement Using Three-Dimensional Dynamic Stochastic Resonance and Classification Based on ResNet

Classification is an important means of extracting rich information from hyperspectral images (HSIs). However, many HSIs contain shadowed areas, where noise severely affects the extraction of useful information. General noise removal may lead to loss of spatial correlation and spectral features. In contrast, dynamic stochastic resonance (DSR) converts noise into capability that enhances the signal in a way that better preserves the image’s original information. Nevertheless, current one-dimensional and 2D DSR methods fail to fully utilize the tensor properties of hyperspectral data and preserve the complete spectral features. Therefore, a hexa-directional differential format is derived in this paper to solve the system’s output, and the iterative equation for HSI shadow enhancement is obtained, enabling 3D parallel processing of HSI spatial–spectral information. Meanwhile, internal parameters are adjusted to achieve optimal resonance. Furthermore, the residual neural network 152 model embedded with the convolutional block attention module is proposed to diminish information redundancy and leverage data concealed within shadow areas. Experimental results on a real-world HSI demonstrate the potential performance of 3D DSR in enhancing weak signals in HSI shadow regions and the proposed approach’s effectiveness in improving classification.

A study on denoising with deep convolutional neural networks in spatial heterodyne spectroscopy

Spatial heterodyne spectroscopy is a hyperspectral remote sensing technology. With the continuous improvement of signal detection accuracy, new methods are urgently needed to further reduce the interference of noise on the information contained in spatial heterodyne interferograms. Convolutional neural networks, as an emerging field, have achieved good results in image denoising in recent years. This paper applies convolutional neural networks to the field of spatial heterodyne spectroscopy, constructs and trains four deep convolutional neural network models for denoising spatial heterodyne interferograms under different brightness and Gaussian noise conditions, and compares their performance with other algorithms. The results show that deep convolutional neural networks have significant advantages in denoising spatial heterodyne interferograms. Finally, a comparison of four neural networks was conducted to explore how to reasonably select network models. This work provides new ideas and effective solutions for reducing the interference of noise on spatial heterodyne spectral information and improving signal detection accuracy.

A Spatial–Spectral Transformer for Hyperspectral Image Classification Based on Global Dependencies of Multi-Scale Features

Vision transformers (ViTs) are increasingly utilized for HSI classification due to their outstanding performance. However, ViTs encounter challenges in capturing global dependencies among objects of varying sizes, and fail to effectively exploit the spatial–spectral information inherent in HSI. In response to this limitation, we propose a novel solution: the multi-scale spatial–spectral transformer (MSST). Within the MSST framework, we introduce a spatial–spectral token generator (SSTG) and a token fusion self-attention (TFSA) module. Serving as the feature extractor for the MSST, the SSTG incorporates a dual-branch multi-dimensional convolutional structure, enabling the extraction of semantic characteristics that encompass spatial–spectral information from HSI and subsequently tokenizing them. TFSA is a multi-head attention module with the ability to encode attention to features across various scales. We integrated TFSA with cross-covariance attention (CCA) to construct the transformer encoder (TE) for the MSST. Utilizing this TE to perform attention modeling on tokens derived from the SSTG, the network effectively simulates global dependencies among multi-scale features in the data, concurrently making optimal use of spatial–spectral information in HSI. Finally, the output of the TE is fed into a linear mapping layer to obtain the classification results. Experiments conducted on three popular public datasets demonstrate that the MSST method achieved higher classification accuracy compared to state-of-the-art (SOTA) methods.

Green Space Reverse Pixel Shuffle Network: Urban Green Space Segmentation Using Reverse Pixel Shuffle for Down-Sampling from High-Resolution Remote Sensing Images

Urban green spaces (UGS) play a crucial role in the urban environmental system by aiding in mitigating the urban heat island effect, promoting sustainable urban development, and ensuring the physical and mental well-being of residents. The utilization of remote sensing imagery enables the real-time surveying and mapping of UGS. By analyzing the spatial distribution and spectral information of a UGS, it can be found that the UGS constitutes a kind of low-rank feature. Thus, the accuracy of the UGS segmentation model is not heavily dependent on the depth of neural networks. On the contrary, emphasizing the preservation of more surface texture features and color information contributes significantly to enhancing the model’s segmentation accuracy. In this paper, we proposed a UGS segmentation model, which was specifically designed according to the unique characteristics of a UGS, named the Green Space Reverse Pixel Shuffle Network (GSRPnet). GSRPnet is a straightforward but effective model, which uses an improved RPS-ResNet as the feature extraction backbone network to enhance its ability to extract UGS features. Experiments conducted on GaoFen-2 remote sensing imagery and the Wuhan Dense Labeling Dataset (WHDLD) demonstrate that, in comparison with other methods, GSRPnet achieves superior results in terms of precision, F1-score, intersection over union, and overall accuracy. It demonstrates smoother edge performance in UGS border regions and excels at identifying discrete small-scale UGS. Meanwhile, the ablation experiments validated the correctness of the hypotheses and methods we proposed in this paper. Additionally, GSRPnet’s parameters are merely 17.999 M, and this effectively demonstrates that the improvement in accuracy of GSRPnet is not only determined by an increase in model parameters.

Self-Supervised Deep Multi-Level Representation Learning Fusion-Based Maximum Entropy Subspace Clustering for Hyperspectral Band Selection

As one of the most important techniques for hyperspectral image dimensionality reduction, band selection has received considerable attention, whereas self-representation subspace clustering-based band selection algorithms have received quite a lot of attention with good effect. However, many of them lack the self-supervision of representations and ignore the multi-level spectral–spatial information of HSI and the connectivity of subspaces. To this end, this paper proposes a novel self-supervised multi-level representation learning fusion-based maximum entropy subspace clustering (MLRLFMESC) method for hyperspectral band selection. Firstly, to learn multi-level spectral–spatial information, self-representation subspace clustering is embedded between the encoder layers of the deep-stacked convolutional autoencoder and its corresponding decoder layers, respectively, as multiple fully connected layers to achieve multi-level representation learning (MLRL). A new auxiliary task is constructed for multi-level representation learning and multi-level self-supervised training to improve its capability of representation. Then, a fusion model is designed to fuse the multi-level spectral–spatial information to obtain a more distinctive coefficient matrix for self-expression, where the maximum entropy regularization (MER) method is employed to promote connectivity and the uniform dense distribution of band elements in each subspace. Finally, subspace clustering is conducted to obtain the final band subset. Experiments have been conducted on three hyperspectral datasets, and the corresponding results show that the proposed MLRLFMESC algorithm significantly outperforms several other band selection methods in classification performance.

Spatial-spectral collaborative attention network for hyperspectral unmixing

In recent years, the transformer architecture has demonstrated exceptional feature extraction capabilities in the field of computer vision (CV). Building on this, our paper aims to fully exploit the potential of the attention in transformers and apply it to the task of hyperspectral unmixing (HU). We propose the Spatial Spectral Collaborative Attention Network (SSCA-Net) model. We obtain spectral information with continuous spatial attributes from HSIs in advance, and input it into SSCA-Net together with HSIs. The improved self-cross attention can collaboratively extract spatial-spectral domain information of HSIs, thereby obtaining more accurate abundance scores. In addition, we conduct ablation experiments to investigate the influence of attention with various configurations on the performance of the unmixing process. The performance of the proposed model is evaluated on three real-world datasets: Samson, Jasper, and Houston, and compared with the performance of FCLSU, GLMM, DAEU, CNNAEU, CyCU, and DHTN algorithms.

Multi-head Global Attention and Spatial Spectral Information Fusion for Remote Sensing Image Compression

Multi-head Global Attention and Spatial Spectral Information Fusion for Remote Sensing Image Compression