Auroral breakup detection in all-sky images by  unsupervised learning

Partamies, Noora; Dol, Bas; Teissier, Vincent; Juusola, Liisa; Syrjäsuo, Mikko; Mulders, Hjalmar

doi:https://doi.org/10.5194/angeo-42-103-2024

Articles | Volume 42, issue 1

https://doi.org/10.5194/angeo-42-103-2024

© Author(s) 2024. This work is distributed under
the Creative Commons Attribution 4.0 License.

https://doi.org/10.5194/angeo-42-103-2024

© Author(s) 2024. This work is distributed under
the Creative Commons Attribution 4.0 License.

Articles | Volume 42, issue 1

Regular paper

|

25 Apr 2024

Regular paper |

| 25 Apr 2024

Auroral breakup detection in all-sky images by unsupervised learning

Noora Partamies, Bas Dol, Vincent Teissier, Liisa Juusola, Mikko Syrjäsuo, and Hjalmar Mulders

Data sets

Auroral images with morphological clusters N. Partamies https://doi.org/10.11582/2023.00132

IMAGE data download IMAGE data https://space.fmi.fi/image/www/index.php

UNIS Keograms KHO keograms http://kho.unis.no/Keograms/keograms.php

Model code and software

Automatic morphological classification of auroral structures (https://github.com/Tadlai/auroral-classification) V. Teissier https://github.com/Tadlai/auroral-classification/blob/main/master_thesis-final.pdf

Short summary

Auroral imaging produces large amounts of image data that can no longer be analyzed by visual inspection. Thus, every step towards automatic analysis tools is crucial. Previously supervised learning methods have been used in auroral physics, with a human expert providing ground truth. However, this ground truth is debatable. We present an unsupervised learning method, which shows promising results in detecting auroral breakups in the all-sky image data.