Search

Background: Recently, sparse unmixing has been successfully applied to spectral mixture analysis of remotely sensed hyperspectral images. Based on the assumption that the observed image signatures can be expressed in the form of linear combinations of a number of pure spectral signatures known in advance, unmixing of each mixed pixel in the scene is to find an optimal subset of signatures in a very large spectral library, which is cast into the framework of sparse regression. However, traditional sparse regression models, such as collaborative sparse regression, ignore the intrinsic geometric structure in the hyperspectral data. Results: In this paper, we propose a novel model, called manifold regularized collaborative sparse regression, by introducing a manifold regularization to the collaborative sparse regression model. The manifold regularization utilizes a graph Laplacian to incorporate the locally geometrical structure of the hyperspectral data. An algorithm based on alternating direction method of multipliers has been developed for the manifold regularized collaborative sparse regression model. Conclusions: Experimental results on both the simulated and real hyperspectral data sets have demonstrated the effectiveness of our proposed model. [ABSTRACT FROM AUTHOR]

Classification of hyperspectral images always suffers from high dimensionality and very limited labeled samples. Recently, the spectral-spatial classification has attracted considerable attention and can achieve higher classification accuracy and smoother classification maps. In this paper, a novel spectral-spatial classification method for hyperspectral images by using kernel methods is investigated. For a given hyperspectral image, the principle component analysis (PCA) transform is first performed. Then, the first principle component of the input image is segmented into non-overlapping homogeneous regions by using the entropy rate superpixel (ERS) algorithm. Next, the local spectral histogram model is applied to each homogeneous region to obtain the corresponding texture features. Because this step is performed within each homogenous region, instead of within a fixed-size image window, the obtained local texture features in the image are more accurate, which can effectively benefit the improvement of classification accuracy. In the following step, a contextual spectral-texture kernel is constructed by combining spectral information in the image and the extracted texture information using the linearity property of the kernel methods. Finally, the classification map is achieved by the support vector machines (SVM) classifier using the proposed spectral-texture kernel. Experiments on two benchmark airborne hyperspectral datasets demonstrate that our method can effectively improve classification accuracies, even though only a very limited training sample is available. Specifically, our method can achieve from 8.26% to 15.1% higher in terms of overall accuracy than the traditional SVM classifier. The performance of our method was further compared to several state-of-the-art classification methods of hyperspectral images using objective quantitative measures and a visual qualitative evaluation. [ABSTRACT FROM AUTHOR]

In this paper, we present a new lossy compression method for hyperspectral images that aims to optimally compress in both spatial and spectral domains and simultaneously minimizes the effect of the compression on linear spectral unmixing performance. To achieve this, a nonnegative Tucker decomposition is applied. This decomposition is a function of three dimension parameters. By employing a link between this decomposition and the linear spectral mixing model, an optimization problem is defined to find the optimal parameters by minimizing the root-mean-square error between the abundance matrices of the original and reconstructed data sets. The resulting optimization problem is solved by a particle swarm optimization algorithm. An approximate method for fast estimation of the free parameters is introduced as well. Our simulation results show that, in comparison with well-known state-of-the-art lossy compression methods, an improved compression and spectral unmixing performance of the reconstructed hyperspectral image is obtained. It is noteworthy to mention that the superiority of our method becomes more apparent as the compression ratio grows. [ABSTRACT FROM AUTHOR]

Abstract P. ginseng is a precious traditional Chinese functional food, which is used for both medicinal and food purposes, and has various effects such as immunomodulation, anti-tumor and anti-oxidation. The growth year of P. ginseng has an important impact on its medicinal and economic values. Fast and nondestructive identification of the growth year of P. ginseng is crucial for its quality evaluation. In this paper, we propose a FC-CNN network that incorporates spectral and spatial features of hyperspectral images to characterize P. ginseng from different growth years. The importance ranking of the spectra was obtained using the random forest method for optimal band selection. Based on the hyperspectral reflectance data of P. ginseng after radiometric calibration and the images of the best five VNIR bands and five SWIR bands selected, the year-by-year identification of P. ginseng age and its identification experiments for food and medicinal purposes were conducted, and the FC-CNN network and its FCNN and CNN branch networks were tested and compared in terms of their effectiveness in the identification of P. ginseng growth years. It has been experimentally verified that the best year-by-year recognition was achieved by utilizing images from five visible and near-infrared important bands and all spectral curves, and the recognition accuracy of food and medicinal use reached 100%. The FC-CNN network is significantly better than its branching model in the effect of edible and medicinal identification. The results show that for P. ginseng growth year identification, VNIR images have much more useful information than SWIR images. Meanwhile, the FC-CNN network utilizing the spectral and spatial features of hyperspectral images is an effective method for the identification of P. ginseng growth year.

Large-scale self-supervised pretraining of deep learning models is known to be critical in several fields, such as language processing, where its has led to significant breakthroughs. Indeed, it is often more impactful than architectural designs. However, the use of self-supervised pretraining lags behind in several domains, such as hyperspectral images, due to data scarcity. This paper addresses the challenge of data scarcity in the development of methods for spatial super-resolution of hyperspectral images (HSI-SR). We show that state-of-the-art HSI-SR methods are severely bottlenecked by the small paired datasets that are publicly available, also leading to unreliable assessment of the architectural merits of the models. We propose to capitalize on the abundance of high resolution (HR) RGB images to develop a self-supervised pretraining approach that significantly improves the quality of HSI-SR models. In particular, we leverage advances in spectral reconstruction methods to create a vast dataset with high spatial resolution and plausible spectra from RGB images, to be used for pretraining HSI-SR methods. Experimental results, conducted across multiple datasets, report large gains for state-of-the-art HSI-SR methods when pretrained according to the proposed procedure, and also highlight the unreliability of ranking methods when training on small datasets.

Multiscale Dense Network (MSDN) is a deep convolutional neural network method proposed to solve the loss of fine features in hyperspectral remote sensing images. Multiscale Dense Network (MSDN), which makes full use of the different scale information in the network, implements feature extraction of HSI in both horizontal and vertical dimensions. Since MSDN maintains all fine scales until the last layer of the network, the computational effort is large. To address this problem, this paper proposes a lightweight Inter layer Short Multiscale Dense Network (IMSDN) method for hyperspectral image classification. HSI classification is performed using network reduction and inert computation to avoid unnecessary computation, the MSDN network is divided into S blocks along the horizontal dimension, keeping the coarsest (S-i+1) scale only in the ith block and only down sampling the final output of each block in the Short Multiscale Dense Network (SMSDN) for HSI classification. On the basis of the SMSDN network, in order to maintain the classification accuracy of the network, the Lightweight Interlayer Multiscale Dense Network (IMSDN) utilizes three scales of convolution kernels in different horizontal directions to extract the image features separately, which are then fused for the classification of HSI. Experiments are conducted on two publicly available datasets and the results are compared with other methods to demonstrate that the IMSDN network has less computation and shortens the experiment time compared to the MSDN network while guaranteeing the classification accuracy.

It has been proved that hyperspectral image (HSI) classification task benefits from introducing additional spatial information. However, how to classify high-dimensional hyperspectral images under the condition of limited training samples is still a challenge. In this paper, we propose a hyperspectral image classification framework based on discriminant manifold broad learning system (DMBLS). We introduce the manifold structure and discriminant information of the data samples as prior knowledge into broad learning system (BLS), which effectively solves the problem of insufficient learning caused by BLS with limited training samples. Firstly we construct two different types of manifold structures for class modelling: the intra-class manifold structures and the inter-class ones. Secondly, we integrate these local discriminant manifold structures to build a manifold regularization framework. Finally, we add this framework to the DMBLS via minimizing the intra-class manifold structures, while maximizing the inter-class ones. And we think this operation can optimize the projection direction of output weights in the DMBLS, enhancing the discrimination ability of these weights. Experimental results on three benchmark HSI datasets show that DMBLS can effectively improve the classification accuracy of hyperspectral images under the condition of limited training samples. [ABSTRACT FROM AUTHOR]

Hyperspectral images (HSIs) have attracted worldwide attention as they make it possible to remotely distinguish between spectrally similar materials. The data of hyperspectral image with high dimensionality, redundant features, unbalanced distribution and so on, however, impose great difficulties for HSI classification. An ensemble of classifiers are usually considered to provide more accurate predictions than a single classifier. In this paper, a heterogeneous ensemble consisting of a support vector machine, a kernel extreme learning machine and a multinomial logistic regression classifier is constructed to increase classification accuracies of HSIs. Furthermore, the feature extraction method is employed to transform the spectral features to promote the diversity of the ensemble, and a voting mechanism is introduced to predict the label on unseen data. Our experimental results with two widely used real hyperspectral datasets demonstrate that the proposed heterogeneous ensemble classifier obtains satisfactory performance with respect to spectral feature sets, even though the number of labelled samples is small. [ABSTRACT FROM AUTHOR]

• Solved the lack of spatial features of HSI in predicting soil carbon content. • Proposed multi-scale spectral feature network to extract spectral feature of HSI. • Proposed multi-scale spatial and spectral feature network to extract feature of HSI. • Proposed methods improve the accuracy and stability of soil carbon content prediction. Soil carbon content prediction based on hyperspectral images can achieve large-scale spatial measurement, which has the advantages of wide coverage and fast information collection, is more suitable for field data collection. However, the research on soil carbon content prediction based on hyperspectral images mainly focuses on feature extraction of spectral information, ignoring the spatial information, and cannot well reveal the intrinsic structural characteristics of data. Aiming at the lack of spatial features consideration in hyperspectral images, soil carbon content prediction methods based on multi-scale feature fusion are proposed by hyperspectral image. At the same time of extracting spectral features from hyperspectral images, the spatial information is used for the first time and a multi-scale spectral and spatial feature network (SpeSpaMN) is designed. In the SpeSpaMN, the multi-scale spectral feature network (SpeMN) is constructed to extract spectral features, the multi-scale spatial feature network (SpaMN) is constructed to extract spatial features. The two networks are fused by using the complementary relationship between different scale features to achieve soil carbon content prediction based on multi-scale feature fusion. The results showed that SpeSpaMN had the best results compared to other methods, followed by the method of SpeMN. The RPD of Inland, Aoshan Bay and Jiaozhou Bay samples based on SpeSpaMN were increased by 47.36%, 37.96% and 4.30% respectively. This paper can effectively solve the problem of the deep fusion of spatial and spectral features in the soil carbon content prediction by hyperspectral image, so as to improve the accuracy and stability of soil carbon content prediction. [ABSTRACT FROM AUTHOR]

Powdery mildew is considered to be one of the important diseases that threaten wheat production. It can cause serious economic losses and damage food security. Therefore, early detection of powdery mildew infection is essential to limit its amplification in reproduction. The purpose of this study is to use the hyperspectral image (HSI) to detect wheat powdery mildew (WPM) in an objective and accurate way. A method is proposed to classify the infection degree of WPM by combining cascade interval selection sensitive band with support vector machine (SVM). Firstly, HSI of leaf and canopy scale are obtained, and the image was segmented based on HSV (hue, saturation, value) color space and OTSU algorithm (HSV + OTSU). The spectral reflectance data were extracted after removing the background region of interest (ROI), and then the multiple scattering correction (MSC) was performed. The spectral data after MSC was called RAW, and then the first derivative (D1) and second derivative (D2) operations were performed respectively. Subsequently, the method of backward interval partial least squares (BiPLS) was used to screen out the characteristic bands of the sensitive interval. Finally, the classification models of WPM disease degree at the leaf scale and the canopy scale (RAW-BiPLS-SVM, D1-BiPLS-SVM and D2-BiPLS-SVM) were established respectively. The results showed that the spectral data after derivative operation was enhanced, and the classification effect of D1-BiPLS-SVM model had advantages at both leaf and canopy scales, in which the overall accuracy (OA) at the leaf scale was 95.2 %, and the OA at canopy scale was 74.4 %. In order to increase the difference between diseases, the canopy-scale disease severity levels were integrated. The correct rates of health and disease were 96.9 % and 99.2 %, respectively, and the OA of disease grade recombination was 87.9 %. Therefore, the method proposed in this paper can not only accurately identified healthy and infected wheat, but also realized the classification of different degrees of WPM, and provide technical support for the prevention and control of WPM. • Wheat powdery mildew (WPM) was detected based on the hyperspectral data. • A method for WPM infection classification based on interval selection sensitivity bands combined with SVM is proposed. • The proposed method can effectively classify WPM from leaf to canopy levels. • The results of the proposed method suggested its potential in field applications. [ABSTRACT FROM AUTHOR]

Hyperspectral image classification can be achieved by modeling an energy minimization problem on a graph of image pixels. In this paper, an effective spectral-spatial classification method for hyperspectral images based on joint bilateral filtering (JBF) and graph cut segmentation is proposed. In this method, a novel technique for labeling regions obtained by the spectral-spatial segmentation process is presented. Our method includes the following steps. First, the probabilistic support vector machines (SVM) classifier is used to estimate probabilities belonging to each information class. Second, an extended JBF is employed to perform image smoothing on the probability maps. By using our JBF process, salt-and-pepper classification noise in homogeneous regions can be effectively smoothed out while object boundaries in the original image are better preserved as well. Third, a sequence of modified bi-labeling graph cut models is constructed for each information class to extract the desirable object belonging to the corresponding class from the smoothed probability maps. Finally, a classification map is achieved by merging the segmentation maps obtained in the last step using a simple and effective rule. Experimental results based on three benchmark airborne hyperspectral datasets with different resolutions and contexts demonstrate that our method can achieve 8.56%-13.68% higher overall accuracies than the pixel-wise SVM classifier. The performance of our method was further compared to several classical hyperspectral image classification methods using objective quantitative measures and a visual qualitative evaluation. [ABSTRACT FROM AUTHOR]

Due to the complementary properties of different features, multiple feature fusion has a large potential for hyperspectral imagery classification. At the meantime, hashing is promising in representing a high-dimensional float-type feature with extremely low bit binary codes while maintaining the performance. In this paper, we study the possibility of using hashing to fuse multiple features for hyperspectral imagery classification. For this purpose, we propose a multiple feature fusion framework to evaluate the performance of using different hashing methods. For comparison and completeness, we also have an extensive comparison to five subspace-based dimension reduction methods and six fusion-based methods which are popular solutions to deal with multiple features in hyperspectral image classification. Experimental results on four benchmark hyperspectral data sets demonstrate that using hashing to fuse multiple features can achieve comparable or better performance with the traditional subspace-based dimension reduction methods and fusion-based methods. Moreover, the binary features obtained by using hashing need much less storage and are faster to compute distances with the help of machine instructions. [ABSTRACT FROM PUBLISHER]

The algebraic multigrid (AMG) method is used to solve linear systems of equations on a series of progressively coarser grids and has recently attracted significant attention for image segmentation due to its high efficiency and robustness. In this paper, a novel spectral-spatial classification method for hyperspectral images based on the AMG method and hierarchical segmentation (HSEG) algorithm is proposed. Our method consists of the following steps. First, the AMG method is applied to hyperspectral imagery to construct a multigrid structure of fine-to-coarse grids based on the anisotropic diffusion partial differential equation (PDE). The vertices in the multigrid structure are then considered as the initial seeds (markers) for growing regions and are clustered to obtain a sequence of segmentation results. In the next step, a maximum vote decision rule is employed to combine the pixel-wise classification map and the segmentation maps. Finally, a final classification map is produced by choosing the optimal grid level to extract representative spectra. Experiments based on three different types of real hyperspectral datasets with different resolutions and contexts demonstrate that our method can obtain 3.84%-13.81% higher overall accuracies than the SVM classifier. The performance of our method was further compared to several marker-based spectral-spatial classification methods using objective quantitative measures and a visual qualitative evaluation. [ABSTRACT FROM AUTHOR]

Hyperspectral images are widely used in several real-life applications. In this paper, we investigate on the compression of hyperspectral images by considering different aspects, including the optimization of the computational complexity in order to allow implementations on limited hardware (i.e., hyperspectral sensors, etc.). We present an approach that relies on a three-dimensional predictive structure. Our predictive structure, 3D-MBLP, uses one or more previous bands as references to exploit the redundancies among the third dimension. The achieved results are comparable, and often better, with respect to the other state-of-art lossless compression techniques for hyperspectral images. [ABSTRACT FROM AUTHOR]

Augmenting spectral data with spatial information for image classification has recently gained significant attention, as classification accuracy can often be improved by extracting spatial information from neighboring pixels. In this paper, we propose a new framework in which active learning (AL) and hierarchical segmentation (HSeg) are combined for spectral–spatial classification of hyperspectral images. The spatial information is extracted from a best segmentation obtained by pruning the HSeg tree using a new supervised strategy. The best segmentation is updated at each iteration of the AL process, thus taking advantage of informative labeled samples provided by the user. The proposed strategy incorporates spatial information in two ways: 1) concatenating the extracted spatial features and the original spectral features into a stacked vector and 2) extending the training set using a self-learning-based semi-supervised learning (SSL) approach. Finally, the two strategies are combined within an AL framework. The proposed framework is validated with two benchmark hyperspectral datasets. Higher classification accuracies are obtained by the proposed framework with respect to five other state-of-the-art spectral–spatial classification approaches. Moreover, the effectiveness of the proposed pruning strategy is also demonstrated relative to the approaches based on a fixed segmentation. [ABSTRACT FROM PUBLISHER]

Dimensionality reduction is a common approach to decrease the high computational complexity and improve the classification performance of hyperspectral images. The paper addresses this issue by introducing a novel semisupervised clustering approach to hyperspectral image classification. In this approach, the relationships between the samples (i.e., pixels in hyperspectral data) are measured by two kinds of side constraints, i.e., positive and negative constraints, which are imposed to construct a discriminative transformation that establishes a regularized metric function. Accordingly, a new subspace is built in which the class discrimination capability of each individual feature is expanded, while the spectral correlation among features is greatly reduced. Then, the learned metric is formulated within an exemplar-based clustering framework, i.e., the affinity propagation (AP). Thus, the proposed approach is called decorrelation–separability-based AP (DS-AP). Experimental results obtained on three hyperspectral remote sensing data sets demonstrate the effectiveness of the proposed DS-AP technique for hyperspectral image classification. [ABSTRACT FROM PUBLISHER]

Dictionary learning and sparse coding have been widely studied as mechanisms for unsupervised feature learning. Unsupervised learning could bring enormous benefit to the processing of hyperspectral images and to other remote sensing data analysis because labelled data are often scarce in this field. We propose a method for clustering the pixels of hyperspectral images using sparse coefficients computed from a representative dictionary as features. We show empirically that the proposed method works more effectively than clustering on the original pixels. We also demonstrate that our approach, in certain circumstances, outperforms the clustering results of features extracted using principal component analysis and non-negative matrix factorisation. Furthermore, our method is suitable for applications in repetitively clustering an ever-growing amount of high-dimensional data, which is the case when working with hyperspectral satellite imagery. [ABSTRACT FROM AUTHOR]

In this paper, a novel adaptive algorithm for spectral unmixing in hyperspectral images (HSIs) is proposed. Many of the existing spectral unmixing algorithms, under the assumption of the linear model for the spectral mixing phenomenon, attempt to estimate the signatures of available materials in the observed HSI image. Then, based on the similarity between the estimated spectral signatures and the available spectral signatures in the spectral library, they identify the materials in the HSI and estimate their relative abundances. While the spectral library, as prior knowledge, has not been directly considered in the founding of existing algorithms, the proposed method is directly concentrated on the spectral signatures library. Assuming the linear spectral mixing model, the proposed method takes a set of spectral signatures which are probably present in the observed HSI. Then, based on a non-statistical approach, the normalized least mean square (NLMS) adaptive algorithm is engaged to estimate a weight vector for each spectral signature in the selected set in such a way that each weight vector and its corresponding spectral signature are non-orthogonal whereas the weight vector of each spectral signature is almost orthogonal to the other spectral signatures. A synthetic dataset of hyperspectral images is considered to evaluate the performance of the proposed method. The evaluation results show that the proposed method outperforms its counterparts in low signal to noise ratio (SNR). [ABSTRACT FROM AUTHOR]

Precise and timely classification of land cover types plays an important role in land resources planning and management. In this paper, nine kinds of land cover types in the acquired hyperspectral scene are classified based on the kernel collaborative representation method. To reduce the spectral shift caused by adjacency effect when mining the spatial-spectral features, a correlation coefficient-weighted spatial filtering operation is proposed in this paper. Additionally, by introducing this operation into the kernel collaborative representation method with Tikhonov regularization (KCRT) and discriminative KCRT (DKCRT) method, respectively, the weighted spatial-spectral KCRT (WSSKCRT) and weighted spatial-spectral DKCRT (WSSDKCRT) methods are constructed for land cover classification. Furthermore, aiming at the problem of difficulty of pixel labeling in hyperspectral images, this paper attempts to establish an effective land cover classification model in the case of small-size labeled samples. The proposed WSSKCRT and WSSDKCRT methods are compared with four methods, i.e., KCRT, DKCRT, KCRT with composite kernel (KCRT-CK), and joint DKCRT (JDKCRT). The experimental results show that the proposed WSSKCRT method achieves the best classification performance, and WSSKCRT and WSSDKCRT outperform KCRT-CK and JDKCRT, respectively, obtaining the OA over 94% with only 540 labeled training samples, which indicates that the proposed weighted spatial filtering operation can effectively alleviate the spectral shift caused by adjacency effect, and it can effectively classify land cover types under the situation of small-size labeled samples.

Hyperspectral images (HSIs) are unavoidably degraded by mixed noise, including Gaussian noise and sparse noise. In this paper, we consider a constrained tubal rank and sparsity model (CTSD) to tackle the HSIs mixed noise removal, which characterizes the clean HSIs via the low-tubal-rank constraint and the sparse noise via the l 0 and l ∞ norm constraints, respectively. Due to the strong non-convexity, the CTSD model is challenging to solve. To tackle the CTSD, we develop the proximal alternating minimization (PAM) algorithm via the exact tensor singular value decomposition (t-SVD) and establish the global convergence under mild assumptions. Since the t-SVD is computationally expensive, especially for large scale images, we further design an efficient inexact PAM algorithm via an inexact t-SVD. The inexact PAM enjoys two advantages: (1) The computational complexity for SVDs of the inexact PAM (O (rn 1 n 2 n 3)) is about twofold faster than that of the exact PAM (O (min (n 1 , n 2) n 1 n 2 n 3)) for r ≪ min (n 1 , n 2) ; (2) The accuracy of the inexact PAM is theoretically guaranteed. Extensive experiments on HSIs denoising demonstrate that the exact and inexact methods both outperform comparing methods in quantitative evaluation metrics and visual effects, and the inexact PAM can compromise between the accuracy and efficiency for large scale HSIs. [ABSTRACT FROM AUTHOR]

The vision transformer (ViT) has demonstrated performance comparable to that of convolutional neural networks (CNN) in the hyperspectral image classification domain. This is achieved by transforming images into sequence data and mining global spectral-spatial information to establish remote dependencies. Nevertheless, both the ViT and CNNs have their own limitations. For instance, a CNN is constrained by the extent of its receptive field, which prevents it from fully exploiting global spatial-spectral features. Conversely, the ViT is prone to excessive distraction during the feature extraction process. To be able to overcome the problem of insufficient feature information extraction caused using by a single paradigm, this paper proposes an MLP-mixer and a graph convolutional enhanced transformer (MGCET), whose network consists of a spatial-spectral extraction block (SSEB), an MLP-mixer, and a graph convolutional enhanced transformer (GCET). First, spatial-spectral features are extracted using SSEB, and then local spatial-spectral features are fused with global spatial-spectral features by the MLP-mixer. Finally, graph convolution is embedded in multi-head self-attention (MHSA) to mine spatial relationships and similarity between pixels, which further improves the modeling capability of the model. Correlation experiments were conducted on four different HSI datasets. The MGEET algorithm achieved overall accuracies (OAs) of 95.45%, 97.57%, 98.05%, and 98.52% on these datasets.

Although the collaborative use of hyperspectral images (HSIs) and LiDAR data in land cover classification tasks has demonstrated significant importance and potential, several challenges remain. Notably, the heterogeneity in cross-modal information integration presents a major obstacle. Furthermore, most existing research relies heavily on category names, neglecting the rich contextual information from language descriptions. Visual-language pretraining (VLP) has achieved notable success in image recognition within natural domains by using multimodal information to enhance training efficiency and effectiveness. VLP has also shown great potential for land cover classification in remote sensing. This paper introduces a dual-sensor multimodal semantic collaborative classification network (DSMSC2N). It uses large language models (LLMs) in an instruction-driven manner to generate land cover category descriptions enriched with domain-specific knowledge in remote sensing. This approach aims to guide the model to accurately focus on and extract key features. Simultaneously, we integrate and optimize the complementary relationship between HSI and LiDAR data, enhancing the separability of land cover categories and improving classification accuracy. We conduct comprehensive experiments on benchmark datasets like Houston 2013, Trento, and MUUFL Gulfport, validating DSMSC2N’s effectiveness compared to various baseline methods.

Extreme learning machine (ELM) is an efficient learning algorithm that has been recently applied to hyperspectral image classification. In this paper, the first implementation of the ELM algorithm fully developed for graphical processing unit (GPU) is presented. ELM can be expressed in terms of matrix operations so as to take advantage of the single instruction multiple data (SIMD) computing paradigm of the GPU architecture. Additionally, several techniques like the use of ensembles, a spatial regularization algorithm, and a spectral–spatial classification scheme are applied and projected to GPU in order to improve the accuracy results of the ELM classifier. In the last case, the spatial processing is based on the segmentation of the hyperspectral image through a watershed transform. The experiments are performed on remote sensing data for land cover applications achieving competitive accuracy results compared to analogous support vector machine (SVM) strategies with significantly lower execution times. The best accuracy results are obtained with the spectral–spatial scheme based on applying watershed and a spatially regularized ELM. [ABSTRACT FROM PUBLISHER]

This paper explores the performance of Ensembles of Extreme Learning Machine classifiers for hyperspectral image classification and segmentation in a semisupervised and spatially regularized process. The approach assumes that we have available only a small training set of labeled samples, which we enrich with a set of guessed labelings on selected samples from the vast pool of unlabeled image pixels. Selection and label guessing is conditioned to an unsupervised classification of the image pixel spectra, and to the spatial proximity to the labeled samples in the image domain. Unlabeled pixels falling in the spatial neighborhood of a labeled training sample, and belonging to the same unsupervised class, acquire its label. Unsupervised classification can be performed by any clustering technique, in this paper we have resorted to the classical K-means. The classifier built from the enriched training dataset is applied to the entire hyperspectral image. Finally, we perform a spatial regularization of the classification label image, maximizing a rather general prior smoothness criterion, by the selection of the most frequent class in each pixel neighborhood. This paper reports experiments with homogeneous ensembles of ELM, rELM, and OP-ELM classifiers, including a sensitivity analysis over the ensemble size and the number of hidden nodes. Computational experiments on four well known benchmarking hyperspectral images give state-of-the-art results. [ABSTRACT FROM AUTHOR]

Because there are many unlabeled samples in hyperspectral images and the cost of manual labeling is high, this paper adopts semi-supervised learning method to make full use of many unlabeled samples. In addition, those hyperspectral images contain much spectral information and the convolutional neural networks have great ability in representation learning. This paper proposes a novel semi-supervised hyperspectral image classification framework which utilizes self-training to gradually assign highly confident pseudo labels to unlabeled samples by clustering and employs spatial constraints to regulate self-training process. Spatial constraints are introduced to exploit the spatial consistency within the image to correct and re-assign the mistakenly classified pseudo labels. Through the process of self-training, the sample points of high confidence are gradually increase, and they are added to the corresponding semantic classes, which makes semantic constraints gradually enhanced. At the same time, the increase in high confidence pseudo labels also contributes to regional consistency within hyperspectral images, which highlights the role of spatial constraints and improves the HSIc efficiency. Extensive experiments in HSIc demonstrate the effectiveness, robustness, and high accuracy of our approach. [ABSTRACT FROM AUTHOR]

Hyperspectral Images are a challenging area because of big dimensions, spatial and spectral low resolution. This paper aims to propose a method to provide an accurate and quick classification based on tensor decomposition and map segmentation. For reducing the dimensions of hyperspectral data and obtain the abundances map we used Parafac decomposition. For determining the segmentation abundances map representing the segmentation abundances matrix, an algorithm based on the Dynamic Clustering Algorithm [1] principle was used having as input the estimated abundances matrix. The segmentation abundances map was visually labeled using the true color image. [ABSTRACT FROM AUTHOR]

We propose a new learning rule for sparse multinomial logistic regression (SMLR). The new rule is the generalization of the one proposed in the pioneering work by Krishnapuram et al. In our proposed method, the parameters of SMLR are iteratively estimated from log-posterior by using some approximations. The proposed update rule provides a faster convergence compared to the state-of the-art methods used for SMLR parameter estimation. The estimated parameters are tested on the pixel-based classification of hyperspectral images. The experimental results on real hyperspectral images show that the classification accuracy of proposed method is also better than those of the state-of-the-art methods. [ABSTRACT FROM AUTHOR]

This paper proposes a whale optimized subspace matching pursuit algorithm to improve the precision of sparse unmixing of hyperspectral data. The proposed algorithm modifies the traditional whale group optimization algorithm by introducing the nonlinear population control parameters and evolution strategy and thus improves the convergence speed and convergence precision of the whale cluster optimization algorithm. Based on the improved subspace matching pursuit algorithm, the new algorithm uses the whale optimization algorithm to solve the abundance coefficients of the known end members, where the objective function is modeled as a constrained sparse regression. In order to improve the precision of the subspace matching pursuit algorithm, the minimum reconstruction error is taken as the criterion and the redundant end members with smaller coefficients are removed to the greatest extent. The simulation and real data experiments show that the proposed algorithm can effectively remove the redundant endmembers and has a higher precision of unmixing. [ABSTRACT FROM AUTHOR]

• Observing the fact that the 2-D snapshot measurement can preserve certain spatial structure the HSI, we use the snapshot measurement to group nonlocal similar cubes of the unknown HSI, which is both effective and efficient. • A tensor-based reconstruction algorithm is developed using nonlocal low-rank tensor train and 3-D weighted total variation that can effectively characterize the nonlocal self-similarity and the local smoothness properties of the HSI. • A series of experiments on both simulated and real CASSI datasets are carried out to verify the outperformance of NLTT-TV over several other state-of-the-art methods. In this paper, we consider the reconstruction problem of hyperspectral images (HSIs) based on a new type of imaging system termed as coded aperture snapshot spectral imaging (CASSI). CASSI compresses multiple spectral bands of HSI into a single snapshot measurement which enjoys low cost, low bandwidth, and high-speed sensing rate. Despite the promising applications of CASSI systems, the bottleneck lies in the reconstruction algorithms. The existing CASSI reconstruction methods could not fully exploit the underlying structures of HSIs, and thereby demonstrate unsatisfactory reconstruction quality. To remedy this issue, we propose in this paper a novel tensor-based method by modeling two intrinsic HSI priors, namely nonlocal low-rank tensor train (NLTT) and 3-D weighted total variation (3DWTV). Taking into consideration that the snapshot measurement and targeted HSI should share similar spatial structures, we use the snapshot measurement to help group nonlocal spatially similar cubes within the unknown HSI to form 4-D tensors. And the NLTT prior employed on these 4-D tensors can effectively learn the correlations among the spatial, spectral and nonlocal modes thanks to the well-balanced matricization scheme of TT rank. Meanwhile, we employ the 3DWTV prior to characterize the local smoothness of the spatial and spectral modes. The resulting model can be solved efficiently by using the alternative direction method of multipliers (ADMM). Extensive experiments on both simulated and real CASSI datasets validate the improved reconstruction performance of our method over several other competing ones. [ABSTRACT FROM AUTHOR]

Band selection is a frequently used dimension reduction technique for hyperspectral images (HSI) to address the "curse of dimensionality" phenomenon in machine learning (ML). This technique identifies and selects a subset of the most important bands from the original ones to remove redundancy and noisy information while maintaining optimal generalization ability. Band selection methods can be categorized into supervised and unsupervised techniques depending on whether labels are used or not. An unsupervised band selection and feature extraction framework is proposed in this study. The framework trains a sub-neural network to identify the most important and informative bands from the original data space, which is then projected to a reduced and more informative space. The classification performance of the selected bands' combination on the Pavia University HSI datasets has been verified using multiple machine learning algorithms. The proposed method not only enhances the classification results of HSI, but also reduces the computational time and data storage requirements compared to other state-of-the-art band selection approaches. [ABSTRACT FROM AUTHOR]

Recently, the autoencoder (AE) based method plays a critical role in the hyperspectral anomaly detection domain. However, due to the strong generalised capacity of AE, the abnormal samples are usually reconstructed well along with the normal background samples. Thus, in order to separate anomalies from the background by calculating reconstruction errors, it can be greatly beneficial to reduce the AE capability for abnormal sample reconstruction while maintaining the background reconstruction performance. A memory‐augmented autoencoder for hyperspectral anomaly detection (MAENet) is proposed to address this challenging problem. Specifically, the proposed MAENet mainly consists of an encoder, a memory module, and a decoder. First, the encoder transforms the original hyperspectral data into the low‐dimensional latent representation. Then, the latent representation is utilised to retrieve the most relevant matrix items in the memory matrix, and the retrieved matrix items will be used to replace the latent representation from the encoder. Finally, the decoder is used to reconstruct the input hyperspectral data using the retrieved memory items. With this strategy, the background can still be reconstructed well while the abnormal samples cannot. Experiments conducted on five real hyperspectral anomaly data sets demonstrate the superiority of the proposed method. [ABSTRACT FROM AUTHOR]

Because of the complicated imaging conditions in space and the finite imaging systems on satellites, the resolution of remote sensing images is limited. The process of increasing an image's resolution, image super-resolution, aims to obtain a clearer image. High-resolution (HR) images are affected by various input conditions, such as motion, imaging blur, down-sampling matrix, and various types of noise. Changes in these conditions seriously affect low-resolution (LR) images, so if the imaging process is a pathological problem, super-resolution reconstruction is a pathological anti-problem. To optimize the imaging quality of satellites without changing the optical system, we chose to reconstruct images acquired by satellites using deep learning. We changed the original super-resolution generative adversarial nets network, upgraded the generator's network part to ResNet-50, and inserted an additional fully connected (FC) layer in the network of the discriminator part. We also modified the loss function by changing the weight of regularization loss from 2 × 10−8 to 2 × 10−9, aiming to preserve more detail. In addition, we carefully and specifically chose remote sensing images taken under low-light circumstances from GF-5 satellites to form a new dataset for training and validation. The test results proved that our method can obtain good results. The reconstruction peak signal-to-noise ratio (PSNR) at the scaling factors of 2, 3, and 4 reached 32.6847, 31.8191, and 30.5095 dB, respectively, and the corresponding structural similarity (SSIM) reached 0.8962, 0.8434, and 0.8124. The super-resolution speed was also satisfactory, making real-time reconstruction more probable. [ABSTRACT FROM AUTHOR]

The deep network model relies on sufficient training samples to achieve superior processing performance, which limits its application in hyperspectral image (HSI) classification. In order to perform HSI classification with noisy labels, a robust weakly supervised feature learning (WSFL) architecture combined with multi-model attention is proposed. Specifically, the input noisy labeled data are first subjected to multiple groups of residual spectral attention models and multi-granularity residual spatial attention models, enabling WSFL to refine and optimize the extracted spectral and spatial features, with a focus on extracting clean samples information and reducing the model's dependence on labels. Finally, the fused and optimized spectral-spatial features are mapped to the multilayer perceptron (MLP) classifier to increase the constraint of the model on the noisy samples. The experimental results on public datasets, including Pavia Center, WHU-Hi LongKou, and HangZhou, show that WSFL is better at classifying noise labels than excellent models such as spectral-spatial residual network (SSRN) and dual channel residual network (DCRN). On Hangzhou dataset, the classification accuracy of WSFL is superior to DCRN by 6.02% and SSRN by 7.85%, respectively. [ABSTRACT FROM AUTHOR]

Hyperspectral imaging instruments could capture detailed spatial information and rich spectral signs of observed scenes. Much spatial information and spectral signatures of hyperspectral images (HSIs) present greater potential for detecting and classifying fine crops. The accurate classification of crop kinds utilizing hyperspectral remote sensing imaging (RSI) has become an indispensable application in the agricultural domain. It is significant for the prediction and growth monitoring of crop yields. Amongst the deep learning (DL) techniques, Convolution Neural Network (CNN) was the best method for classifying HSI for their incredible local contextual modeling ability, enabling spectral and spatial feature extraction. This article designs a Hybrid Multi-Strategy Aquila Optimization with a Deep Learning-Driven Crop Type Classification (HMAODL-CTC) algorithm onHSI. The proposed HMAODL-CTC model mainly intends to categorize different types of crops on HSI. To accomplish this, the presented HMAODL-CTC model initially carries out image preprocessing to improve image quality. In addition, the presented HMAODL-CTC model develops dilated convolutional neural network (CNN) for feature extraction. For hyperparameter tuning of the dilated CNN model, the HMAO algorithm is utilized. Eventually, the presented HMAODL-CTC model uses an extreme learning machine (ELM) model for crop type classification. A comprehensive set of simulations were performed to illustrate the enhanced performance of the presented HMAODL-CTC algorithm. Extensive comparison studies reported the improved performance of the presented HMAODL-CTC algorithm over other compared methods. [ABSTRACT FROM AUTHOR]

Soil carbon content prediction based on hyperspectral images can achieve large-scale spatial measurement, which has the advantages of wide coverage and fast information collection, is more suitable for field data collection. However, the research on soil carbon content prediction based on hyperspectral images mainly focuses on feature extraction of spectral information, ignoring the spatial information, and cannot well reveal the intrinsic structural characteristics of data. Aiming at the lack of spatial features consideration in hyperspectral images, soil carbon content prediction methods based on multi-scale feature fusion are proposed by hyperspectral image. At the same time of extracting spectral features from hyperspectral images, the spatial information is used for the first time and a multi-scale spectral and spatial feature network (SpeSpaMN) is designed. In the SpeSpaMN, the multi-scale spectral feature network (SpeMN) is constructed to extract spectral features, the multi-scale spatial feature network (SpaMN) is constructed to extract spatial features. The two networks are fused by using the complementary relationship between different scale features to achieve soil carbon content prediction based on multi-scale feature fusion. The results showed that SpeSpaMN had the best results compared to other methods, followed by the method of SpeMN. The RPD of Inland, Aoshan Bay and Jiaozhou Bay samples based on SpeSpaMN were increased by 47.36%, 37.96% and 4.30% respectively. This paper can effectively solve the problem of the deep fusion of spatial and spectral features in the soil carbon content prediction by hyperspectral image, so as to improve the accuracy and stability of soil carbon content prediction.

Endmember extraction is the process of selecting pure spectral signatures of materials from hyperspectral data. Most of the endmember extraction methods in the literature use only the spectral information, and disregard the spatial composition of the image. Spatial-spectral preprocessing methods, motivated by the assumption that endmembers are more likely to be located in homogenous regions, can increase the performance of endmember extraction by directing the extraction process to homogenous regions. However, such an approach generally results in a failure of extracting anomalous or scarce endmembers, which can be important in practical applications, e.g., to extract endmembers of materials such as landmines, rare minerals, or stressed crops. Although anomaly detection can be applied in parallel to endmember extraction, the process of endmember extraction and unmixing provides a summary of the data, which is important for concepts such as data scanning and compression, and disregarding anomalous endmembers in such a summary or compression of big data may result in undesired consequences for many application fields. In this paper, an approach that guides the endmember extraction process to spatially homogenous regions instead of transition areas, while also extracting anomalous pixel vectors as endmembers, is proposed. The proposed approach can be used with any spectral-based endmember extraction method. The experimental results validate the approach for both synthetic and real hyperspectral images. [ABSTRACT FROM PUBLISHER]

Copyright of Arabian Journal for Science & Engineering (Springer Science & Business Media B.V. ) is the property of Springer Nature and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

Abstract: This paper focuses on the application of Extreme Learning Machines (ELM) to the classification of remote sensing hyperspectral data. The specific aim of the work is to obtain accurate thematic maps of soybean crops, which have proven to be difficult to identify by automated procedures. The classification process carried out is as follows: First, spectral data is transformed into a hyper-spherical representation. Second, a robust image gradient is computed over the hyper-spherical representation allowing an image segmentation that identifies major crop plots. Third, feature selection is achieved by a greedy wrapper approach. Finally, a classifier is trained and tested on the selected image pixel features. The classifiers used for feature selection and final classification are Single Layer Feedforward Networks (SLFN) trained with either the ELM or the incremental OP-ELM. Original image pixel features are computed following a Functional Data Analysis (FDA) characterization of the spectral data. Conventional ELM training of the SLFN improves over the classification performance of state of the art algorithms reported in the literature dealing with the data treated in this paper. Moreover, SLFN-ELM uses less features than the referred algorithms. OP-ELM is able to find competitive results using the FDA features from a single spectral band. [Copyright &y& Elsevier]

This paper introduces the Hybrid Extreme Rotation Forest (HERF) classifier describing two succesful applications in the image segmentation domain. The HERF is an ensemble of classifiers composed of Extreme Learning Machines (ELM) and Decision Trees. Training of the HERF includes optimal rotation of random partitions of the feature set aimed to increase diversity. The first application is the segmentation of 3D Computed Tomography Angiography (CTA) following an Active Learning (AL) strategy for the optimal sample selection to minimize the number of data samples needed to obtain a required accuracy degree. AL is pertinent for interactive learning processes where a human operator is required to select training samples to enhance the classifier in an iterative process, therefore labeling samples for training may be a time consuming and expensive process. CTA image segmentation is one of such processes, due to the variability in CTA images which hinders the generalization of classifiers trained on one dataset to new datasets. Following an AL strategy, the human operator is presented with a visual selection of pixels whose labeling would be most informative for the classifier. After adding those labeled pixels to the training data, the classifier is retrained. This iteration is repeated until image segmentation quality meets the required level. The approach is applied to the segmentation of the thrombus in CTA imaging of Abdominal Aortic Aneurysm (AAA) patients, showing that the structures of interest can be accurately segmented after a few iterations using a small data sample. The second application is a new semisupervised segmentation algorithm for hyperspectral images. The algorithm steps are: 1) supervised training an initial classifier from a small balanced training set, 2) clustering of the image pixels, by a k-means algorithm 3) adding unlabeled pixels to the original trainning data set according to the spatial neighborhood and the cluster membership, 4) supervised training of the classifier with the enriched training data set, 6) classification of the hyperspectral image 4) spatial regularization of classification results consisting in selecting the most frequent class in each pixel neighborhood. Results on two well known benchmarking hyperspectral images improve over state of the art algorithms. [ABSTRACT FROM AUTHOR]

Hyperspectral images represent an important source of information to assess ecosystem biodiversity. In particular, plant species richness is a primary indicator of biodiversity. This paper uses spectral variance to predict vegetation richness, known as Spectral Variation Hypothesis. Hierarchical agglomerative clustering is our primary tool to retrieve clusters whose Shannon entropy should reflect species richness on a given zone. However, in a high spectral mixing scenario, an additional unmixing step, just before entropy computation, is required; cluster centroids are enough for the unmixing process. Entropies computed using the proposed method correlate well with the ones calculated directly from synthetic and field data. [ABSTRACT FROM AUTHOR]