Hyperspectral Image Clustering Research Articles

Contrastive Multiview Subspace Clustering of Hyperspectral Images Based on Graph Convolutional Networks

Contrastive Multiview Subspace Clustering of Hyperspectral Images Based on Graph Convolutional Networks

Superpixel-Based and Spatially Regularized Diffusion Learning for Unsupervised Hyperspectral Image Clustering

Superpixel-Based and Spatially Regularized Diffusion Learning for Unsupervised Hyperspectral Image Clustering

SDST: Self-Supervised Double-Structure Transformer for Hyperspectral Images Clustering

SDST: Self-Supervised Double-Structure Transformer for Hyperspectral Images Clustering

From Model-Based Optimization Algorithms to Deep Learning Models for Clustering Hyperspectral Images

Hyperspectral images (HSIs), captured by different Earth observation airborne and space-borne systems, provide rich spectral information in hundreds of bands, enabling far better discrimination between ground materials that are often indistinguishable in visible and multi-spectral images. Clustering of HSIs, which aims to unveil class patterns in an unsupervised way, is highly important in the interpretation of HSI, especially when labelled data are not available. A number of HSI clustering methods have been proposed. Among them, model-based optimization algorithms, which learn the cluster structure of data by solving convex/non-convex optimization problems, have achieved the current state-of-the-art performance. Recent works extend the model-based algorithms to deep versions with deep neural networks, obtaining huge breakthroughs in clustering performance. However, a systematic survey on the topic is absent. This article provides a comprehensive overview of clustering methods of HSI and tracked the latest techniques and breakthroughs in the domain, including the traditional model-based optimization algorithms and the emerging deep learning based clustering methods. With a new taxonomy, we elaborated on the main ideas, technical details, advantages, and disadvantages of different types of clustering methods of HSIs. We provided a systematic performance comparison between different clustering methods by conducting extensive experiments on real HSIs. Unsolved problems and future research trends in the domain are pointed out. Moreover, we provided a toolbox that contains implementations of representative clustering algorithms to help researchers to develop their own models.

Fast low-rank-matrix-based connection center evolution for large hyperspectral image clustering

Clustering is a common data processing technique and has been increasingly applied in remote sensing processing. In recent years, a novel clustering algorithm called connection center evolution (CCE) has been proposed. To a certain extent, CCE solves the parameter sensitivity problem, which generally exists in traditional clustering algorithms. However, CCE is not feasible for handling large hyperspectral images (HSI) due to its computational bottlenecks. That is, when the size of HSI to be processed is large, the construction of the similarity matrix and the calculation of its power have very high space complexity and time complexity, respectively. To solve this problem, we propose a novel approach for large HSI datasets, named fast low-rank-matrix-based connectivity center evolution (LMCCE). The proposed LMCCE algorithm first introduces the KNN graph into CCE to construct a sparse similarity matrix, effectively reducing its space complexity. Then, symmetric non-negative matrix factorization (SymNMF) is used for the similarity matrix, which can maintain most of the information while reducing the redundancy and noise in similarity matrix. Furthermore, by performing SymNMF on the similarity matrix, the problem of calculating the power of the similarity matrix is transformed into the problem of calculating the power of a small matrix, which significantly reduces the time complexity. Experimental results on several synthetic and real HSI datasets demonstrate that the proposed LMCCE algorithm is faster than CCE, and can handle large HSI while achieving superior performance compared to state-of-the-art clustering methods.

Recognition of spider mite infestations in jujube trees based on spectral-spatial clustering of hyperspectral images from UAVs.

Spider mite infestations are a serious hazard for jujube trees in China. The use of remote sensing technology to evaluate the health of jujube trees in large-scale intensive agricultural production is an effective means of agricultural control. Hyperspectral remote sensing has a higher spectral resolution and richer spectral information than conventional multispectral remote sensing, which improves the detection of crop pests and diseases. We used hyperspectral remote sensing data from jujube fields infested with spider mite in Hotan Prefecture, Xinjiang to evaluate their use in monitoring this important pest. We fused spectral and spatial information from the hyperspectral images and propose a method of recognizing spider mite infestations of jujube trees. Our method is based on the construction of spectral features, the fusion of spatial information and clustering of these spectral-spatial features. We evaluated the effect of different spectral-spatial features and different clustering methods on the recognition of spider mite in jujube trees. The experimental results showed that the overall accuracy of the method for the recognition of spider mites was >93% and the overall accuracy of the band clustering-density peak clustering model for the recognition of spider mite reached 96.13%. This method can be applied to the control of jujube spider mites in agricultural production.

Robust spectral clustering integrated nonlocal information blocks for hyperspectral image classification

Hyperspectral image (HSI) clustering is a challenging task due to its containing rich spectral information and spatial features. Subspace clustering has been used to explore the intrinsic connections between data points successfully. Graph convolutional subspace clustering (GCSC) was used to achieve robust HSI clustering. The model remaps the self-expressions of the data to non-Euclidean domains to generate a robust graph embedding dictionary. The efficient kernel graph convolutional subspace clustering model utilizes a subspace clustering model with the Frobenius norm and a Gaussian kernel function to achieve a globally optimal closed-form solution that is easy to implement, train, and apply. To reduce the impact of image noise on the segmentation effect, we consider spatial nonlocal self-similarity using nonlocal information blocks based on the fact that the spectral features of HSIs lie in low-dimensional subspaces. At the same time, the three-dimensional tensor generated by the nonlocal similar image blocks is estimated using a robust statistic function. Principal component analysis is used to remove the redundant information of hyperspectral data and extract the main features of pixels. Then a robust statistical function and nonlocal information block are used to construct the affinity matrix. The redundant information prevalent in the whole natural image is used for denoising, that is, the image block is taken as the unit to find similar areas in the image, then these regions are weighted and averaged. Since all pixels in the image are used, the noise can be better filtered. Finally, the constructed affinity matrix is applied to spectral clustering to obtain better clustering results. The model used is called nonlocal mean robust statistical kernel graph convolutional subspace clustering (N-RKGCSC). Experimental results on Salinas, Indian Pines, Pavia Center, and Pavia University datasets showed that the N-RKGCSC model could eliminate the interference of additive and multiplicative noise well, achieving better segmentation performance and better noise immunity.

Graph-Based Structural Deep Spectral-Spatial Clustering for Hyperspectral Image

As a fundamental task in hyperspectral image (HSI) applications, HSI clustering has been extensively researched over the years. Inspired by the capability of graphs in structure modeling, some recent works have exploited the local graph-structured information among pixels to boost the HSI clustering performance. However, these works only explore shallow graph structures for representation learning, while ignoring the beneficial high-order as well as global structure characteristics of HSI pixels, thus resulting in limited clustering performance. In this paper, a novel graph-based structural deep spectral-spatial clustering network (GSDSSC) is proposed to sufficiently explore the structure information among pixels. Specifically, a self-expression embedded multi-graph auto-encoder is proposed to explore high-order structure associations among pixels, thereby capturing robust spectral-spatial features and global clustering structure. Besides, to learn more compact spectral-spatial features for clustering, a global structure-guided optimization mechanism is further designed to constrain the local feature distribution with global clustering structure. Extensive experiments on three public HSI benchmark datasets illustrate the superior performance of the proposed GSDSSC over state-of-the-art HSI clustering methods.

Semi-Supervised Neural Architecture Search for Hyperspectral Imagery Classification Method With Dynamic Feature Clustering

Hyperspectral image(HSI) contains rich spatial and spectral information, which makes HSI classification task the research focus of HSI analysis within remote sensing community. Though deep learning based HSI classification methods obtain good performance in recent years, how to learn network structure better suitable for a given HSI instead of utilizing a manually designed one for HSI classification is still a challenging problem, especially providing only small amount of labeled samples. To address this problem, we propose the first semi-supervised HSI classification network constructed via the neural architecture search. Specifically, we propose a two-head semi-supervised HSI classification framework utilizing both labeled and unlabeled data, which consists of a shared feature extraction module, a classifier module for labeled samples together with a clustering module for unlabeled samples. To boost the performance of the constructed two-head network, we propose to utilize deep features instead of the original pixels for HSI clustering to generate pseudo labels for the unlabeled data. Within the conducted semi-supervised network, we specifically design a method to automatically search for the shared feature extraction module better suitable for the given HSI data, which leads to better HSI classification results. Experimental results on three HSI datasets demonstrate the effectiveness of the proposed method, providing only limited number of labeled training samples.

Deep Spectral Clustering With Regularized Linear Embedding for Hyperspectral Image Clustering

The past decade has witnessed the rapid development of deep learning techniques, especially for large-scale and complex data sets. However, it is still a noteworthy problem in dealing with unsupervised hyperspectral image segmentation since inefficiency and misleading result from the absence of supervised information. Generally, spectral clustering is one of the most powerful clustering algorithms, as it often outperforms other methods for image segmentation. Unfortunately, the poor scalability and generalization severely limit the use of spectral clustering, especially for large-scale and high-dimensional hyperspectral images processing. The major motivation of this work is to solve this problem, and we designed a novel algorithm, termed Deep Spectral Clustering with Regularized Linear Embedding (DSCRLE), to benefit from both spectral graph theory and deep learning techniques. The brief procedure is first to construct a fully connected neural network to extract latent feature representations, and then normalize the feature representations by the spectral orthonormal constraint. Lastly, by introducing low-dimensional embedding, we refined the final outputs of all given unlabeled hyperspectral pixels. Extensive experiments have demonstrated that the competitiveness of the proposed method, and it outperforms state-of-art clustering approaches in the task of hyperspectral image segmentation.

Clustering information-constrained 3D U-Net subspace clustering for hyperspectral image

ABSTRACT Hyperspectral image (HSI) clustering is a challenging task due to the complex spatial-spectral structure and high-dimensional property in HSI data. In this letter, a novel clustering information-constrained 3D U-Net subspace clustering network is proposed for HSI clustering. Considering the spatial-spectral information, the proposed network takes the 3D pixel cubes around the pixels as the input. Based on the 3D pixel cubes, a 3D U-Net subspace clustering network is introduced to extract spatial-spectral features from 3D pixel cubes and learn self-representation subspace property among pixels. In order to learn features more suitable for clustering, a clustering information constraint is introduced to explore useful information gain in the existing clustering result. Experiments conducted on three public HSI datasets illustrate the superior performance of the proposed method.

Deep Convolutional Asymmetric Autoencoder-Based Spatial-Spectral Clustering Network for Hyperspectral Image

Due to the complex properties of hyperspectral images (HSI), such as spatial-spectral structure, high dimension, and great spectral variability, HSI clustering is a challenging operation. In this paper, we propose a novel deep convolutional asymmetric autoencoder-based spatial-spectral clustering network (DCAAES2C-Net) which employs a convolutional autoencoder (CAE) and an asymmetric autoencoder to investigate spatial-spectral information. First, we use a CAE to extract spatial-spectral features. Then, we introduce an asymmetric autoencoder between the encoder and decoder of CAE to suppress some non-material-related spatial information in latten feature like shading and texture. By using a collaborative strategy to train the proposed networks, we obtain the representation features in a low dimension. Furthermore, we improve the k -means algorithm by using the concept of over-clustering to handle fuzzy representation which is difficult to distinguish the cluster, and utilize it to obtain the final HSI clustering result. The results of the experiments demonstrated that the proposed methodology outperforms other methods on the frequently used hyperspectral image dataset.

Spatial-Spectral Constrained Adaptive Graph for Hyperspectral Image Clustering.

Hyperspectral image (HSI) clustering is a challenging task, whose purpose is to assign each pixel to a corresponding cluster. The high-dimensionality and noise corruption are two main problems that limit the performance of HSI clustering. To address those problems, this paper proposes a projected clustering with a spatial–spectral constrained adaptive graph (PCSSCAG) method for HSI clustering. PCSSCAG first constructs an adaptive adjacency graph to capture the accurate local geometric structure of HSI data adaptively. Then, a spatial–spectral constraint is employed to simultaneously explore the spatial and spectral information for reducing the negative influence on graph construction caused by noise. Finally, projection learning is integrated into the spatial–spectral constrained adaptive graph construction for reducing the redundancy and alleviating the computational cost. In addition, an alternating iteration algorithm is designed to solve the proposed model, and its computational complexity is theoretically analyzed. Experiments on two different scales of HSI datasets are conducted to evaluate the performance of PCSSCAG. The associated experimental results demonstrate the superiority of the proposed method for HSI clustering.

Fast robust fuzzy clustering based on bipartite graph for hyper‐spectral image classification

Hyper-spectral image (HSI) clustering has become a hot spot in remote sensing image research. However, due to the large amount and high dimension of HSI data, some traditional image clustering methods are no longer suitable, and even lead to poor clustering performance and long computing time. Therefore, this paper proposes a novel clustering method for HSI to make up for the shortcomings of traditional clustering methods. Firstly, a fuzzy similarity matrix is constructed by using the bipartite graph to obtain low-dimensional hyper-spectral data, which can reduce the complexity and the operation time of the algorithm. Secondly, fuzzy membership mapping is carried out through the constructed bipartite graph, and a non-negative regularization term is added to limit ill-conditioned problems. Thirdly, this paper uses the Geman-McClure function to optimize Euclidean distance. Finally, in this paper, the spectral information of HSI is considered while the neighbourhood information of HSI is added to achieve the unity of space spectrum, which enhances the robustness of the algorithm and improves the clustering performance of the algorithm. Compared with existing HSI clustering algorithms, the proposed algorithm has higher clustering accuracy and efficiency.

Robust Spectral Clustering Incorporating Statistical Sub-Graph Affinity Model

Hyperspectral image (HSI) clustering is a challenging work due to its high complexity. Subspace clustering has been proven to successfully excavate the intrinsic relationships between data points, while traditional subspace clustering methods ignore the inherent structural information between data points. This study uses graph convolutional subspace clustering (GCSC) for robust HSI clustering. The model remaps the self-expression of the data to non-Euclidean domains, which can generate a robust graph embedding dictionary. The EKGCSC model can achieve a globally optimal closed-form solution by using a subspace clustering model with the Frobenius norm and a Gaussian kernel function, making it easier to implement, train, and apply. However, the presence of noise can have a noteworthy negative impact on the segmentation performance. To diminish the impact of image noise, the concept of sub-graph affinity is introduced, where each node in the primary graph is modeled as a sub-graph describing the neighborhood around the node. A statistical sub-graph affinity matrix is then constructed based on the statistical relationships between sub-graphs of connected nodes in the primary graph, thus counteracting the uncertainty image noise by using more information. The model used in this work was named statistical sub-graph affinity kernel graph convolutional subspace clustering (SSAKGCSC). Experiment results on Salinas, Indian Pines, Pavia Center, and Pavia University data sets showed that the SSAKGCSC model can achieve improved segmentation performance and better noise resistance ability.

Spectral clustering with anchor graph based on set-to-set distances for large-scale hyperspectral images

ABSTRACT Since labelled samples of hyperspectral images (HSIs) may be unavailable in practical remote sensing applications, large-scale HSI clustering is very important. Due to the huge amount of data brought by the rich spectral and spatial information in large-scale HSIs, HSI clustering is still a challenging task. Among the methods designed for large-scale HSIs clustering, the anchor graph-based methods simultaneously inherit the merits of graph-based clustering and reduce the computational complexity by introducing anchor samples to graph construction. However, the affinity matrix computed by inaccurate distances between anchor samples and other HSI samples can hardly obtain satisfactory clustering performance. To solve this problem, we propose a novel approach for large-scale HSI clustering, namely, spectral clustering with anchor graph based on set-to-set distances (SCAG-SSD) derived from local covariance matrix representation (LCMR). First, superpixels and LCMR features of HSI are obtained via the entropy rate superpixel algorithm and maximum noise fraction, respectively. Second, pure and anomalous samples of each superpixel are distinguished via the distances of LCMR features, and then anchor samples are selected via statistics of the distances between pure samples in each superpixel. In this way, selected anchor samples are representative enough to link all the HSI samples. Third, pure samples in each superpixel, anomalous and anchor samples with their corresponding nearest neighbouring samples are all regarded as different sets. The set-to-set distance is achieved by weighting the LCMR-based distances between samples of two sets. Finally, fast anchor graph clustering is conducted based on set-to-set distances to obtain final clustering maps. Extensive experiments conducted on three publicly available benchmark HSIs demonstrate that the proposed method achieves state-of-the-art clustering accuracy with comparable efficiency.

Affinity Propagation Based on Structural Similarity Index and Local Outlier Factor for Hyperspectral Image Clustering

In hyperspectral remote sensing, the clustering technique is an important issue of concern. Affinity propagation is a widely used clustering algorithm. However, the complex structure of the hyperspectral image (HSI) dataset presents challenge for the application of affinity propagation. In this paper, an improved version of affinity propagation based on complex wavelet structural similarity index and local outlier factor is proposed specifically for the HSI dataset. In the proposed algorithm, the complex wavelet structural similarity index is used to calculate the spatial similarity of HSI pixels. Meanwhile, the calculation strategy of the spatial similarity is simplified to reduce the computational complexity. The spatial similarity and the traditional spectral similarity of the HSI pixels jointly constitute the similarity matrix of affinity propagation. Furthermore, the local outlier factors are applied as weights to revise the original exemplar preferences of the affinity propagation. Finally, the modified similarity matrix and exemplar preferences are applied, and the clustering index is obtained by the traditional affinity propagation. Extensive experiments were conducted on three HSI datasets, and the results demonstrate that the proposed method can improve the performance of the traditional affinity propagation and provide competitive clustering results among the competitors.

Self-Supervised Learning of Discriminative Spatial–Spectral Features for Hyperspectral Images Clustering

Deep subspace clustering methods have demonstrated its outstanding capability for hyperspectral image (HSI) clustering. However, there is a lack of explicit supervision to ensure that low-dimensional spatial–spectral features learned from high-dimensional HSI cubes have good subspace structures. In this letter, we propose a cascade residual capsule network (CRCN) for extracting deep spatial–spectral features and introduce the coding rate reduction (CRR) to measure the compactness of learned spatial–spectral features. We exploit the difference between the coding rate of all features and the sum of that of features of each category as loss function, which provides explicit supervision for learning invariant spatial–spectral features to HSI cube flips and rotations and facilitate the subsequent HSI clustering. To learn features that are discriminative to diverse spatial context information on land-cover objects of the same category, we add a regularization term to the loss function, which is a brink loss between the <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">${l_{2}}$ </tex-math></inline-formula> -norm of active vectors of class capsules in the proposed CRCN and the assigned labels. Experimental results on three benchmark HSI datasets demonstrate the effectiveness of the proposed method, which achieves superiority performance over several state-of-the-art HSI clustering methods.

MS$^{2}$A-Net: Multiscale Spectral–Spatial Association Network for Hyperspectral Image Clustering

Remote sensing hyperspectral cameras acquire high spectral-resolution data that reveal valuable composition information on the targets (e.g., for Earth observation and environmental applications). The intrinsic high dimensionality and the lack of sufficient numbers of labeled/training samples prevent efficient processing of hyperspectral images (HSIs). HSI clustering can alleviate these limitations. In this study, we propose a multi-scale spectral-spatial association network (MS <inline-formula><tex-math notation="LaTeX">$^{2}$</tex-math></inline-formula> A-Net) to cluster HSIs. The backbone of MS <inline-formula><tex-math notation="LaTeX">$^{2}$</tex-math></inline-formula> A-Net is an autoencoder architecture that allows the network to capture the non-linear relation between data points in an unsupervised manner. The network applies a multi-stream approach. One stream extracts spectral information by deploying a spectral association unit. The other stream derives multi-scale contextual and spatial information by employing dilated (atrous) convolutional kernels. The obtained feature representation generated by MS <inline-formula><tex-math notation="LaTeX">$^{2}$</tex-math></inline-formula> A-Net is fed into a standard k-means clustering algorithm to produce the final clustering result. Extensive experiments on four HSIs for different types of applications (i.e., geological-, rural-, and urban-mapping) demonstrate the superior performance of MS <inline-formula><tex-math notation="LaTeX">$^{2}$</tex-math></inline-formula> A-Net over the state-of-the-art shallow/deep learning-based clustering approaches in terms of clustering accuracy. The code is available at:&#x00A0;<uri>https://github.com/Kasra2020/MS2A-Net</uri>.

A Bipartite Graph Partition-Based Coclustering Approach With Graph Nonnegative Matrix Factorization for Large Hyperspectral Images

Clustering large hyperspectral images (HSIs) is a very challenging problem because large HSIs have high dimensionality, large spectral variability, and large computational and memory consumption. Recently, sparse subspace clustering (SSC) has achieved remarkable success in HSI clustering. However, most SSC-based methods suffer from the following bottlenecks for large HSIs: 1) high computational consumption and memory space during the construction of the similarity matrix and decomposition of the graph Laplacian matrix and 2) failure to capture the relationships among dictionary atoms, sparse coefficients, and hyperspectral pixels. To address these challenges, we propose a novel algorithm that extends SSC to cocluster large HSIs, called bipartite graph partition with graph nonnegative matrix factorization (BGP-GNMF). Specifically, to fully explore the characteristics of the spectral and spatial contexts in HSIs, we propose a novel superpixel and pixel coclustering framework with bipartite graph partitioning in the joint sparse representation domain, where superpixel-based dictionary atoms are defined as disjoint vertex sets of the bipartite graph and the joint sparsity representation is mapped into the adjacency matrix of the undirected bipartite graph. To overcome the challenges of high computational consumption and large memory space for large HSIs, the bipartite graph partition with orthonormal constrained nonnegative matrix factorization is proposed to simultaneously cluster the structured dictionary atoms and hyperspectral pixels with an indicator matrix. Finally, to exploit the intrinsic geometry of HSIs, we incorporate manifold regularization into the bipartite graph partition to improve final clustering accuracy. The effectiveness and efficiency of the proposed method are verified on three classical HSIs, and the experimental results illustrate the superiority of the proposed method compared with other state-of-the-art HSI clustering methods.