Open AccessSelection of features system and network parameters for hyperspectral images classification using convolutional neural networks
Author(s) -
V. I. Kozik,
E. S. Nezhevenko
Publication year - 2021
Language(s) - English
Resource type - Conference proceedings
DOI - 10.25743/sdm.2021.28.23.020
Subject(s) - hyperspectral imaging , convolutional neural network , artificial intelligence , pattern recognition (psychology) , computer science , contextual image classification , artificial neural network , image (mathematics) , selection (genetic algorithm) , dimension (graph theory) , mathematics , pure mathematics
A classification system for hyperspectral images using convolutional neural networks is described. A specific network was selected and analyzed. The network parameters, ensured the maximum classification accuracy: dimension of the input layer, number of the layers, size of the fragments into which the classified image is divided, number of learning epochs, are experimentally determined. High percentages of correct classification were obtained with a large-format hyperspectral image, and some of the classes into which the image is divided are very close to each other and, accordingly, are difficult to distinguish by hyperspectra.