This paper deals with hyperspectral image classification in remote sensing. The proposed scheme is a spectral-spatial technique based on wavelet transforms and mathematical morphology. The original contribution of this paper is that the extended morphological profile (EMP) is created from the features extracted by wavelets, which has proven to be better or comparable to other techniques for dimensionality reduction of hyperspectral data. In addition, the hyperspectral image is denoised, also using wavelets, with the objective of removing undesirable artifacts introduced in the acquisition of the data. The classification is carried out by a support vector machine (SVM) classifier. The accuracy is improved when comparing with previously developed spectral-spatial SVM-based schemes.

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.

Among the different computational intelligence techniques avalaible for hyperspectral data classification, support vector machines (SVMs) have played a dominant role. Recently, a new learning algorithm for single-layer feedforward neural networks called the extreme learning machine (ELM) was proposed. This technique is competitive with SVMs in terms of accuracy, learning speed, and computational scalability. In this article, we propose and evaluate the use of ELM for land-cover classification from hyperspectral images. In addition, we consider two ELM-based techniques integrating spectral and spatial information of the image. The first is a scheme that uses a majority vote approach in order to combine the results of a pixel-wise spectral classification by ELM and a segmentation map obtained by a watershed algorithm. The second introduces spatial information from a small spatial neighbourhood after the classification by ELM. We show the usefulness of spatial–spectral ELM-based classification techniques in hyperspectral imaging. The results are compared to those obtained by similar SVM-based techniques and show improved classification results and much lower execution time. These simple and computationally cheap methods can be combined with others traditionally applied to hyperspectral images.

Deep learning (DL) has been shown to obtain superior results for classification tasks in the field of remote sensing hyperspectral imaging. Superpixel-based techniques can be applied to DL, significantly decreasing training and prediction times, but the results are usually far from satisfactory due to overfitting. Data augmentation techniques alleviate the problem by synthetically generating new samples from an existing dataset in order to improve the generalization capabilities of the classification model. In this paper we propose a novel data augmentation framework in the context of superpixel-based DL called dual-window superpixel (DWS). With DWS, data augmentation is performed over patches centered on the superpixels obtained by the application of simple linear iterative clustering (SLIC) superpixel segmentation. DWS is based on dividing the input patches extracted from the superpixels into two regions and independently applying transformations over them. As a result, four different data augmentation techniques are proposed that can be applied to a superpixel-based CNN classification scheme. An extensive comparison in terms of classification accuracy with other data augmentation techniques from the literature using two datasets is also shown. One of the datasets consists of small hyperspectral small scenes commonly found in the literature. The other consists of large multispectral vegetation scenes of river basins. The experimental results show that the proposed approach increases the overall classification accuracy for the selected datasets. In particular, two of the data augmentation techniques introduced, namely, dual-flip and dual-rotate, obtained the best results.