An international team of scientists brought automated space weather prediction a step closer to reality via a neural network that can identify coronal holes—gaps in the solar atmosphere left by particles that cause geomagnetic storms on Earth—in space-based observations.
Robert Jarolim at Austria's University of Graz said CHRONNOS (Coronal Hole RecOgnition Neural Network Over multi-Spectral-data) applies artificial intelligence to spot coronal holes "based on their intensity, shape, and magnetic field properties, which are the same criteria as a human observer takes into account."
The team trained the convolutional neural network on about 1,700 extreme ultraviolet wavelength images of the sun's corona recorded in 2010-2017, and compared its results to 261 manually identified coronal holes.
CHRONNOS matched human performance in 98% of the cases, and outperformed humans in identifying coronal holes from magnetic field maps.
From Skolkovo Institute of Science and Technology (Russia)
View Full Article
Abstracts Copyright © 2021 SmithBucklin, Washington, DC, USA