Deep Learning for Hyperspectral Image Classification: An Overview

Hyperspectral Image (HSI) classification plays a key role remote sensing field. Recently, deep learning has demonstrated its effectiveness in HSI Classification field. This paper presents a spectral-spatial HSI classification technique established on the deep learning based deep belief network (DBN) for deep and abstract feature extraction and adaptive boundary adjustment based segmentation. Proposed approach focuses on integrating the deep learning based spectral features and segmentation based spatial features into a framework for improved performance. Specifically, first the deep DBN model is exploited as a spectral feature extraction based classifier to extract the deep spectral features. Second, spatial contextual features are obtained by utilizing effective adaptive boundary adjustment based segmentation technique. Finally, maximum voting based criteria is operated to integrate the results of extracted spectral and spatial information for improved HSI classification. In general, exploiting spectral features from DBN process and spatial features from segmentation and integration of spectral and spatial information by maximum voting based criteria, has a substantial effect on the performance of HSI classification. Experimental performance on real and widely used hyperspectral data sets with different contexts and resolutions demonstrates the accuracy of the proposed technique and performance is comparable to several recently proposed HSI classification techniques.

Deep neural network has been extensively applied to hyperspectral image (HSI) classification recently. However, its success is greatly attributed to numerous labeled samples, whose acquisition costs a large amount of time and money. In order to improve the classification performance while reducing the labeling cost, this article presents an active deep learning approach for HSI classification, which integrates both active learning and deep learning into a unified framework. First, we train a convolutional neural network (CNN) with a limited number of labeled pixels. Next, we actively select the most informative pixels from the candidate pool for labeling. Then, the CNN is fine-tuned with the new training set constructed by incorporating the newly labeled pixels. This step together with the previous step is iteratively conducted. Finally, Markov random field (MRF) is utilized to enforce class label smoothness to further boost the classification performance. Compared with the other state-of-the-art traditional and deep learning-based HSI classification methods, our proposed approach achieves better performance on three benchmark HSI data sets with significantly fewer labeled samples.

Recently, deep learning-based methods have made great progress in hyperspectral image (HSI) classification (HSIC). Different from ordinary images, the intrinsic complexity of HSIs data still limits the performance of many common convolutional neural network (CNN) models. Thus, the network architecture becomes more and more complex to extract discriminative spectral-spatial features. For instance, 3-D CNN usually has a large number of trainable parameters, thus increasing the computational complexity of the HSIC. In this letter, we designed a cooperative spectral-spatial attention dense network (CS <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> ADN) that takes raw 3-D HSI data as input data. Specifically, the attention module consists of spectral and spatial axes, by which the salient spectral-spatial features will be emphasized. Furthermore, we combined these attention modules with the dense connection, which is termed as the lightweight dense block; it has a lower computation cost and achieves better classification performance. At the same time, we introduced the center loss, by jointly using the supervision of the center loss and the softmax loss, where the discriminative features could be clearly observed, particularly for small data sets. Experimental results on the biased and unbiased HSI data show that our method outperforms several state-of-the-art methods in HSIC with small training samples.

In Hyperspectral Image (HSI) classification, acquiring large quantities of high-quality labeled samples is typically costly and impractical. Traditional deep learning methods are limited in such scenarios due to their dependence on sample quantities. To address this challenge, researchers have turned to Few-Shot Learning (FSL). Although existing FSL methods improve classification performance by enhancing domain invariance through domain adaptation, they often overlook the critical issue of high inter-class similarity and large intra-class variability. Moreover, during domain alignment, features of different categories within the same domain may become confused. To address these issues, this paper proposes a novel Domain-Invariant Few-Shot Contrastive Learning (DIFSCL) method, which combines domain adaptation and contrastive learning strategies to not only learn domain-invariant features but also significantly enhance inter-class discriminability. Based on this, we further design a multi-scale adaptive attention mechanism for a hyperspectral feature extraction network to more effectively extract and optimize generalized features within the DIFSCL framework, significantly improving intra-class consistency and inter-class discriminative capability of features. Experimental results on three widely used HSI datasets demonstrate that our method significantly outperforms existing techniques in few-shot classification tasks.

The lack of training samples remains one of the major obstacles in applying convolutional neural networks (CNNs) to the hyperspectral image (HSI) classification. In this letter, the accurate classification of HSI with limited training samples is investigated. Due to the advantages of minimizing the distance between samples in the same class and maximizing the distance between samples in different classes, siamese CNN is used for HSI classification with limited training samples. After that, to improve the classification performance, data augmentation is investigated for siamese CNN-based HSI classification. Specifically, pair data augmentation based on CutMix is proposed to generate the training pairs of the same or different classes and the new generated training pairs are used to train siamese CNN. Traditional data augmentation methods simply generate new training samples. It is not proper when data augmentation methods keep the loss function and change the content of the input at the same time. Therefore, a soft-loss-based siamese CNN, which changes its loss according to the coupled replacement data augmentation, is proposed to further address the HSI classification with limited training samples. In the experimental part, on two widely used hyperspectral datasets, the influences of different training samples and window sizes are discussed, and the experimental results reveal that the proposed soft-augmentation-based siamese CNN provides competitive results with limited training samples compared with state-of-the-art methods.

Deep learning technologies have been successfully applied to hyperspectral (HS) image classification with remarkable performance. However, compared with traditional machine learning methods, neural networks usually need more data. In remote sensing (RS) research, obtaining a large number of labeled HS data is very difficult and expensive work. Simultaneously, the distribution of feature information is bound to be unbalanced, and tends to conform to the long tail. At present, the neighborhood information of unlabeled samples is usually ignored in HS image classification tasks based on semi-supervised learning. In this letter, we propose a new semi-supervised long-tail learning framework based on spatial neighborhood information (SLN-SNI), which can complete the HS image classification task under unbalanced small sample data. Specifically, a new semi-supervised learning strategy is proposed. On this basis, a new method to determine the label of unlabeled samples based on spatial neighborhood information (SNI) is proposed. The coarse classification results divided into three situations are judged again, and the accuracy of pseudo labels is improved. The performance of the proposed method is tested on three public HS image datasets. Compared with the current advanced methods have achieved a certain improvement.

Hyperspectral image (HSI) classification is a classical task in remote sensing image analysis. With the development of deep learning, schemes based on deep learning have gradually become the mainstream of HSI classification. However, existing HSI classification schemes either lack the exploration of category-specific information in the spectral bands and the intrinsic value of information contained in features at different scales, or are unable to extract multiscale spatial information and global spectral properties simultaneously. To solve these problems, in this article, we propose a novel HSI classification framework named CL-MGNet, which can fully exploit the category-specific properties in spectral bands and obtain features with multiscale spatial information and global spectral properties. Specifically, we first propose a spectral weight learning (SWL) module with a category consistency loss to achieve the enhancement of information in important bands and the mining of category-specific properties. Then, a multiscale backbone is proposed to extract the spatial information at different scales and the cross-channel attention via multiscale convolution and a grouping attention module. Finally, we employ an attention multilayer perceptron (attention-MLP) block to exploit the global spectral properties of HSI, which is helpful for the final fully connected layer to obtain the classification result. The experimental results on five representative hyperspectral remote sensing datasets demonstrate the superiority of our method.

Deep learning techniques have boosted the performance of hyperspectral image (HSI) classification. In particular, convolutional neural networks (CNNs) have shown superior performance to that of the conventional machine learning algorithms. Recently, a novel type of neural networks called capsule networks (CapsNets) was presented to improve the most advanced CNNs. In this paper, we present a modified two-layer CapsNet with limited training samples for HSI classification, which is inspired by the comparability and simplicity of the shallower deep learning models. The presented CapsNet is trained using two real HSI datasets, i.e., the PaviaU (PU) and SalinasA datasets, representing complex and simple datasets, respectively, and which are used to investigate the robustness or representation of every model or classifier. In addition, a comparable paradigm of network architecture design has been proposed for the comparison of CNN and CapsNet. Experiments demonstrate that CapsNet shows better accuracy and convergence behavior for the complex data than the state-of-the-art CNN. For CapsNet using the PU dataset, the Kappa coefficient, overall accuracy, and average accuracy are 0.9456, 95.90%, and 96.27%, respectively, compared to the corresponding values yielded by CNN of 0.9345, 95.11%, and 95.63%. Moreover, we observed that CapsNet has much higher confidence for the predicted probabilities. Subsequently, this finding was analyzed and discussed with probability maps and uncertainty analysis. In terms of the existing literature, CapsNet provides promising results and explicit merits in comparison with CNN and two baseline classifiers, i.e., random forests (RFs) and support vector machines (SVMs).

Most hyperspectral image (HSI) classification methods assume that all classes in the test set are present during training. However, in real-world applications, acquiring labeled training samples is challenging. As a result, it is difficult for the training dataset to cover all possible land cover types, leading to the generalized zero-shot learning (GZSL) problem. Recently, vision-language models (VLMs) have provided rich semantic priors for land cover classes, offering promising potential for GZSL. However, two fundamental gaps hinder their application to HSI classification: the task paradigm gap, arising from the difference between image-level VLMs and the pixel-level HSI classification task; and the knowledge gap, due to the inconsistency between VLM features and HSI spectral–spatial representations. To bridge both gaps, a novel framework leveraging VLM semantic priors for GZSL in HSI classification is proposed, primarily using pseudo-labeling technique to provide knowledge for unseen classes. Specifically, a pseudo-label generation and enhancement module enables a paradigm transition from image-level understanding to pixel-level classification by incorporating HSI’s spatial information. A pseudo-label correction module then refines noisy labels using spectral cues to address the knowledge gap. Finally, a global learning strategy integrates pseudo-label distillation, supervised learning, and feature regularization to classify seen classes while enabling generalization to unseen ones. Experiments on benchmark HSI datasets demonstrate the proposed method’s superiority in generalized zero-shot classification. This work highlights the potential of VLMs in advancing HSI classification in practical applications.

In the research of hyperspectral image classification, deep learning-based methods have achieved promising results. However, due to the high cost and difficulty in obtaining labeled hyperspectral image samples, traditional deep learning methods experience a significant performance decline under few-shot conditions. This paper proposes a deep few-shot learning method, referred to as Wavelet-Based Multi-stream Spatial-Frequency Fusion Network, to address the issue of small sample size in hyperspectral image classification. Through a multi-stream structure, features at various scales are obtained, effectively alleviating the issue of insufficient feature representation caused by limited training samples. These features are then fed into a multi-level wavelet transformation module integrated with a frequency-domain attention module. By decomposing the input into high-frequency and low-frequency components via wavelet transformation, and subsequently inputting these components into the frequency-domain attention module, a richer representation of features in the frequency domain is achieved. Finally, a Spatial-Frequency Fusion module is employed to not only extract abundant feature information but also fuse features from different scales. This process results in feature maps that encompass more comprehensive global feature information, thereby significantly enhancing classification accuracy under small sample conditions. The proposed method has been compared with traditional machine learning approaches and recent deep learning methods on the Houston, Dioni, and University of Pavia datasets. Under training conditions using only 1 to 5, 10, 15, and 20 samples per class, the experimental results demonstrate the superiority of the method proposed in this paper.

Few-shot learning (FSL) uses prior knowledge and supervised experience to effectively classify hyperspectral images (HSIs), thereby reducing the cost of large numbers of labeled samples. However, existing few-shot methods ignore the correlation between cross-domain feature channels, and the feature representation ability is insufficient. To address above issue, this paper proposes a novel Residual Channel Attention Based Sample Adaptation Few-Shot Learning for Hyperspectral Image Classification(RCASA-FSL) for hyperspectral image classification (HSIC), which can capture and enhance cross-domain dependencies through multi-layer residual connection and random-based feature recalibration. Specifically, a Deep Residual Feature Channel Attention Mechanism (DRFCAM) is designed to obtain cross-domain dependencies by residual concatenation, and further the residual structure is stacked for mining depth discrimination information. Furthermore, a new Random-based Feature Recalibration Module (RFRM) is proposed to reassign the feature weights via random matrix, which fully explore feature weight relationships to guide the sample adaptation process. Besides, we design a joint loss function with combining the FSL loss and domain adaptive loss for further optimization model. Experiments conducted on several standard hyperspectral datasets demonstrate that the proposed RCASA-FSL is superior to other FSL techniques in both quantitative and qualitative aspects.

Deep learning models, especially deep convolutional neural networks (CNNs), have been intensively investigated for hyperspectral image (HSI) classification due to their powerful feature extraction ability. In the same manner, ensemble-based learning systems have demonstrated high potential to effectively perform supervised classification. In order to boost the performance of deep learning-based HSI classification, the idea of deep learning ensemble framework is proposed here, which is loosely based on the integration of deep learning model and random subspace-based ensemble learning. Specifically, two deep learning ensemble-based classification methods (i.e., CNN ensemble and deep residual network ensemble) are proposed. CNNs or deep residual networks are used as individual classifiers and random subspaces contribute to diversify the ensemble system in a simple yet effective manner. Moreover, to further improve the classification accuracy, transfer learning is investigated in this study to transfer the learnt weights from one individual classifier to another (i.e., CNNs). This mechanism speeds up the learning stage. Experimental results with widely used hyperspectral datasets indicate that the proposed deep learning ensemble system provides competitive results compared with state-of-the-art methods in terms of classification accuracy. The combination of deep learning and ensemble learning provides a significant potential for reliable HSI classification.

In recent studies, superpixel segmentation has been integrated into hyperspectral (HS) image classification methods. However, the existing superpixel-based classification methods usually suffer from two serious problems. First, the accuracy and efficiency of current superpixel segmentation approaches cannot meet the demands of practical applications for HS images; second, conventional superpixel-based classification methods generally consider each generated superpixel as a unit for the image classification, which may help to reduce the computing time but result in a significant decrease of the classification accuracy. To solve the problems, we propose a fast region growing based superpixel segmentation (FRGSS) algorithm and a novel texture-adaptive superpixel integration strategy (TASIS) for the HS image classification. Experimental results on real Airborne Visible/Infrared Imaging Spectrometer (AVIRIS) HS images demonstrate that the proposed FRGSS outperforms the state-of-the-art superpixel algorithm. In addition, the superiority of the TASIS is verified compared to the pixel-wise and the conventional superpixel-based classification methods.

One of the most important problems in classification of Hyperspectral images is Hughes phenomena. The main reason for this problem is high dimension feature space and deficiency of training samples. To solve this kind of problem, so many methods are applied which some of them emphasize feature reduction and the others point to classification method with low sensitivity to training samples or consider spatial information or semi supervised training samples. In feature reduction algorithms intensity vectors of pixels are used individually. Additionally, in most of such techniques each component of intensity vector is considered as an individual element without attention to inner relation among them. In this paper we consider members of intensity vector as sampling points of spectral reflecting curve (SRC) of corresponding land cover (LC) and use fractal features of SRC in order to classify the picture. Moreover, we will show using these features improve the classification results in multispectral images.

Hyperspectral image (HSI) classification is an important technology for hyperspectral data analysis, which has the challenge of high dimensionality and limited training samples. Moreover, due to the spatial continuity of material distribution, it is reasonable to improve the performance of HSI classification from the local similarity viewpoint, i.e., pixels in the same local region have a high probability to share similar features. Therefore, in this paper, a novel method based on local similarity and kernel low-rank representation is proposed to improve the classification accuracy for HSI. First, taking into consideration that pixels in the same homogeneous region share similar spectral features (termed local similarity), the HSI is segmented by the superpixel segmentation algorithm to obtain the homogeneous regions. Then, a model of kernel low-rank representation based on local similarity (KLRRLS) is proposed for HSI classification. In the model of KLRRLS, the nonlinear structure characteristics of HSI are extracted by employing the kernel trick, and the low-rank constraint is used as a prior to restrict the recovery coefficient matrix in each segmentation region, thus efficiently solving the linear non-separable problem and exploiting the contextual information of HSI. Finally, the classification results are generated by fusing the result of each segmented map via the maximum voting mechanism. Experimental results on three real HSI datasets demonstrate that the proposed method can effectively improve the classification accuracy.