Open AccessNeural network technology to search for targets in remote sensing images of the Earth
Author(s) -
N.S. Abramov,
Alexander Talalaev,
Vitaly Fralenko,
O.G. Shishkin,
Vyacheslav Khachumov
Publication year - 2019
Publication title -
ceur workshop proceedings
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.177
H-Index - 52
ISSN - 1613-0073
DOI - 10.18287/1613-0073-2019-2391-180-186
Subject(s) - computer science , convolutional neural network , artificial intelligence , artificial neural network , class (philosophy) , deep learning , process (computing) , pattern recognition (psychology) , contextual image classification , completeness (order theory) , machine learning , image (mathematics) , mathematical analysis , mathematics , operating system
The paper introduces how multi-class and single-class problems of searching and classifying target objects in remote sensing images of the Earth are solved. To improve the recognition efficiency, the preparation tools for training samples, optimal configuration and use of deep learning neural networks using high-performance computing technologies have been developed. Two types of CNN were used to process ERS images: a convolutional neural network from the nnForge library and a network of the Darknet type. A comparative analysis of the results is obtained. The research showed that the capabilities of convolutional neural networks allow solving simultaneously the problems of searching (localizing) and recognizing objects in ERS images with high accuracy and completeness.