In recent years, different deep learning frameworks were introduced for hyperspectral image (HSI) classification. However, the proposed network models have a higher model complexity, and do not provide high classification accuracy if few-shot learning is used. This paper presents an HSI classification method that combines random patches network (RPNet) and recursive filtering (RF) to obtain informative deep features. The proposed method first convolves image bands with random patches to extract multi-level deep RPNet features. Thereafter, the RPNet feature set is subjected to dimension reduction through principal component analysis (PCA), and the extracted components are filtered using the RF procedure. Finally, the HSI spectral features and the obtained RPNet–RF features are combined to classify the HSI using a support vector machine (SVM) classifier. In order to test the performance of the proposed RPNet–RF method, some experiments were performed on three widely known datasets using a few training samples for each class, and classification results were compared with those obtained by other advanced HSI classification methods adopted for small training samples. The comparison showed that the RPNet–RF classification is characterized by higher values of such evaluation metrics as overall accuracy and Kappa coefficient.
The possibility of using a multifractal approach to study the tectonic disturbance of coals has been investigated. The relationship between the coal disturbance and the asymmetry of fractal spectra of coal images obtained by means of scanning electron microscopy (SEM) is revealed: it has been established that undisturbed coals are characterized, as a rule, by a symmetric fractal dimension spectrum, and the disturbed coals are described by a fractal spectrum with some degree of asymmetry. It is shown that if fractal spectra of images have a symmetric appearance, then brightness distributions of these images are well fitted by a lognormal curves and parameters of these fittings can be estimated through characteristics of the fractal spectra. By using multifractal analysis of images for more than 140 test coal specimens from the quiet zone of a seam and the outburst zone, differences in the brightness distributions for images of coals with various degrees of disturbance were revealed. The basis of the research is the assumption that differences in the structure of disturbed and undisturbed coals are reflected in histograms of the brightness distributions for images of coal specimens. According to the results of multifractal analysis of images for the test coal specimens, it was established that the brightness distributions for images of the surface of undisturbed coal specimens are lognormal, while the brightness distributions for images of the surface of highly disturbed coal specimens, in most cases, deviate from the lognormal one. The conducted studies allow us to conclude about the applicability of the multifractal approach for assessing the degree of coal disturbance using digital images of coal specimens.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.