Single-image super-resolution (SISR) techniques attempt to reconstruct the finer resolution version of a given image from its coarser version. In the SISR of hyperspectral data sets, the simultaneous consideration of spectral bands is crucial for ensuring the spectral fidelity. However, the high spectral resolution of these data sets affects the performance of conventional approaches. This research proposes the design of 3-D convolutional neural network (CNN)-based SISR architectures that can map the spatial-spectral characteristics of hypercubes to a finer spatial resolution. The proposed approaches facilitate the simultaneous optimization of sparse codes and dictionaries with regard to the super-resolution objective. Our main hypothesis is that the consideration of spectral aspects is essential for the spatial enhancement of hyperspectral images. Also, we propose that the regularized deconvolution of a coarser-scale hypercube, using learned 3-D filters, yields the required high-resolution version. Based on these hypotheses, a convolution-deconvolution framework is proposed to super-resolve the hypercubes in parallel with the reconstruction of a set of regularizing features. Novel sparse code optimization sub-networks proposed in this article give better performance than the existing strategies. The endmember similarities and hyperspectral image prior are considered while designing the proposed loss functions. In order to improve the generalizability, a collaborative spectral unmixing strategy is employed to refine the spectral base of the super-resolved result. The spatial-spectral accuracy of the super-resolved hypercubes, in terms of the validity of regularizing features and endmembers, is explored to devise an optimal ensemble strategy. The experiments, over different data sets, confirm better accuracy of the proposed frameworks compared to the prominent approaches.

In this paper, a Capsulenet-based framework is proposed for extracting spectral and spatial features for improving hyperspectral image classification. Unlike conventional strategies, the proposed framework simultaneously optimizes both feature extraction and classification. The spectral features/patterns derived at different levels of hierarchies are remodeled as spectral-feature capsules. Consequently, unlike conventional convolutional neural network-based approaches, the relative locations as well as other properties such as depth, width, and position of the spectral patterns are taken into consideration. In addition to learning spectral features/patterns, a convolutional long short-term memory (conv-LSTM) is employed for sequentially integrating the spatial features learned from each band. The integrated spatial-feature representation, thus obtained from the final hidden state of conv-LSTM, forms spatial-feature capsules. The capsule-level integration of spatial and spectral features/patterns yields better convergence and accuracy as compared to both ensemble-based and kernel-level integrations. Along with the margin loss, a spectral-angle-based reconstruction loss is also minimized to regularize the learning of network weights. Experiments over different standard datasets indicate that the proposed approach performs better than other prominent hyperspectral classifiers. Furthermore, in comparison with the recent deep learning models, our approach is found to be less sensitive to the network parameters and achieves better accuracy even with lesser network depth.

In recent years, deep learning algorithms, particularly convolutional neural networks (CNNs), have significantly improved the performance of the hyperspectral image (HSI) classification. However, due to the high dimensionality of HSI and limited training samples, the deep neural network causes model overfitting. Additionally, considering all the bands of HSI datasets equally for feature learning and being unable to distinguish between the edge and the center pixels of a neighborhood reduces classification accuracy. Thus, in this paper, we propose an end-to-end deep spectral-spatial residual attention network (DSSpRAN) motivated by the attention mechanism of the human visual system for HSI classification. The DSSpRAN considers input HSI data as a 3-D cube instead of using dimensionality reduction methods. The proposed model simultaneously incorporates spectral and spatial features by considering a spectral residual attention network (SRAN) and a spatial residual attention network (SpRAN). In SRAN, the weights are assigned and learned adaptively to select essential features from each band. The SpRAN enhances the importance of classifying each nearby pixel to the center pixel. It assigns the same label as that of the center pixel to the surrounding pixels, thus limiting pixels with different labels. The proposed method has been evaluated on five different datasets to prove the state-of-the-art for various land use land cover scenarios. A comprehensive qualitative and quantitative analysis of the results shows that the proposed method significantly outperforms other state-of-the-art methods.

Salient object detection in hyperspectral images is of interest in various image processing and computer vision applications. Many studies considering spectral information have been reported, extracting only low-level features from a hyperspectral image. This paper proposes a Convolutional Neural Network (CNN) based salient object detection method using hyperspectral imagery to utilise spatial and spectral information simultaneously. The proposed methodology incorporates Extended Morphological Profile (EMP) followed by a CNN to utilise the information from nearby pixels and high-level features simultaneously. We have evaluated the performance of the proposed approach on two independent datasets to verify the generalisation ability, viz. 1) Hyperspectral Salient Object Detection Dataset (HS-SOD) and 2) Pavia University dataset. An extensive quantitative analysis of the results revealed that the proposed method significantly outperforms other state-of-the-art methods by approximately ≥ 2% of AUC (Area Under receiver operating characteristic Curve) and F-measure and lower mean absolute error for both datasets.

Multitask learning is a widely recognized technique in the field of computer vision and deep learning domain. However, it is still a research question in remote sensing, particularly for hyperspectral imaging. Moreover, most of the research in the remote sensing domain focuses on small and single-task-based annotated datasets, which limits the generalizability and scalability of the developed models to more diverse and complex real-world scenarios. Thus, in this study, we propose a multitask deep learning model designed to perform multiple classification and regression tasks simultaneously on hyperspectral images. We validated our approach on a large hyperspectral dataset called TAIGA, which contains 13 forest variables, including three categorical variables and ten continuous variables with different biophysical parameters. We design a sharing encoder and task-specific decoder network to streamline feature learning while allowing each task-specific decoder to focus on the unique aspects of its respective task. Additionally, a dense atrous pyramid pooling layer and attention network were integrated to extract multi-scale contextual information and enable selective information processing by prioritizing task-specific features. Further, we computed multitask loss and optimized its parameters for the proposed framework to improve the model performance and efficiency across diverse tasks. A comprehensive qualitative and quantitative analysis of the results shows that the proposed method significantly outperforms other state-of-the-art methods. We trained our model across 10 seeds/trials to ensure robustness. Our proposed model demonstrates higher mean performance while maintaining lower or equivalent variability. To make the work reproducible, the codes will be available at https://github.com/Koushikey4596/Multitask-Deep-Learning-Model-for-Taiga-datatset.