Auroral Image Classification With Deep Neural Networks
Results from a study of automatic aurora classification using machine learning techniques are presented. The aurora is the manifestation of physical phenomena in the ionosphere-magnetosphere environment. Automatic classification of millions of auroral images from the Arctic and Antarctic is therefor...
Main Authors: | , , , |
---|---|
Other Authors: | , |
Language: | English |
Published: |
Wiley
2022
|
Subjects: | |
Online Access: | https://www.utupub.fi/handle/10024/167395 |
id |
ftunivturku:oai:www.utupub.fi:10024/167395 |
---|---|
record_format |
openpolar |
spelling |
ftunivturku:oai:www.utupub.fi:10024/167395 2023-05-15T13:32:26+02:00 Auroral Image Classification With Deep Neural Networks Andreas Kvammen Derek McKay Noora Partamies Kristoffer Wickstrøm Suomen ESO-keskus, yhteiset, Finnish Centre for Astronomy with ESO - FINCA 2609700 2022-10-28T13:47:07Z https://www.utupub.fi/handle/10024/167395 en eng Wiley Yhdysvallat (USA) United States US 125 e2020JA027808 10.1029/2020JA027808 Journal of Geophysical Research: Space Physics 10 https://www.utupub.fi/handle/10024/167395 URN:NBN:fi-fe2021042823484 2169-9402 2169-9380 2022 ftunivturku 2022-11-03T00:01:44Z Results from a study of automatic aurora classification using machine learning techniques are presented. The aurora is the manifestation of physical phenomena in the ionosphere-magnetosphere environment. Automatic classification of millions of auroral images from the Arctic and Antarctic is therefore an attractive tool for developing auroral statistics and for supporting scientists to study auroral images in an objective, organized, and repeatable manner. Although previous studies have presented tools for detecting aurora, there has been a lack of tools for classifying aurora into subclasses with a high precision ( > 90%). This work considers seven auroral subclasses: breakup, colored, arcs, discrete, patchy, edge, and faint. Six different deep neural network architectures have been tested along with the well-known classification algorithms: k-nearest neighbor (KNN) and a support vector machine (SVM). A set of clean nighttime color auroral images, without clearly ambiguous auroral forms, moonlight, twilight, clouds, and so forth, were used for training and testing the classifiers. The deep neural networks generally outperformed the KNN and SVM methods, and the ResNet-50 architecture achieved the highest performance with an average classification precision of 92%. Other/Unknown Material Antarc* Antarctic Arctic University of Turku: UTUPub Arctic Antarctic |
institution |
Open Polar |
collection |
University of Turku: UTUPub |
op_collection_id |
ftunivturku |
language |
English |
description |
Results from a study of automatic aurora classification using machine learning techniques are presented. The aurora is the manifestation of physical phenomena in the ionosphere-magnetosphere environment. Automatic classification of millions of auroral images from the Arctic and Antarctic is therefore an attractive tool for developing auroral statistics and for supporting scientists to study auroral images in an objective, organized, and repeatable manner. Although previous studies have presented tools for detecting aurora, there has been a lack of tools for classifying aurora into subclasses with a high precision ( > 90%). This work considers seven auroral subclasses: breakup, colored, arcs, discrete, patchy, edge, and faint. Six different deep neural network architectures have been tested along with the well-known classification algorithms: k-nearest neighbor (KNN) and a support vector machine (SVM). A set of clean nighttime color auroral images, without clearly ambiguous auroral forms, moonlight, twilight, clouds, and so forth, were used for training and testing the classifiers. The deep neural networks generally outperformed the KNN and SVM methods, and the ResNet-50 architecture achieved the highest performance with an average classification precision of 92%. |
author2 |
Suomen ESO-keskus, yhteiset, Finnish Centre for Astronomy with ESO - FINCA 2609700 |
author |
Andreas Kvammen Derek McKay Noora Partamies Kristoffer Wickstrøm |
spellingShingle |
Andreas Kvammen Derek McKay Noora Partamies Kristoffer Wickstrøm Auroral Image Classification With Deep Neural Networks |
author_facet |
Andreas Kvammen Derek McKay Noora Partamies Kristoffer Wickstrøm |
author_sort |
Andreas Kvammen |
title |
Auroral Image Classification With Deep Neural Networks |
title_short |
Auroral Image Classification With Deep Neural Networks |
title_full |
Auroral Image Classification With Deep Neural Networks |
title_fullStr |
Auroral Image Classification With Deep Neural Networks |
title_full_unstemmed |
Auroral Image Classification With Deep Neural Networks |
title_sort |
auroral image classification with deep neural networks |
publisher |
Wiley |
publishDate |
2022 |
url |
https://www.utupub.fi/handle/10024/167395 |
geographic |
Arctic Antarctic |
geographic_facet |
Arctic Antarctic |
genre |
Antarc* Antarctic Arctic |
genre_facet |
Antarc* Antarctic Arctic |
op_relation |
125 e2020JA027808 10.1029/2020JA027808 Journal of Geophysical Research: Space Physics 10 https://www.utupub.fi/handle/10024/167395 URN:NBN:fi-fe2021042823484 2169-9402 2169-9380 |
_version_ |
1766027008080347136 |