Multi-channel coronal hole detection with convolutional neural networks^⋆

¹ University of Graz, Institute of Physics, Universitätsplatz 5, 8010 Graz, Austria
e-mail: robert.jarolim@uni-graz.at
² University of Graz, Kanzelhöhe Observatory for Solar and Environmental Research, Kanzelhöhe 19, 9521 Treffen am Ossiacher See, Austria
³ Columbia Astrophysics Laboratory, Columbia University, 550 West 120th Street, New York, NY 10027, USA
⁴ Max-Planck-Institut für Sonnensystemforschung, Justus-von-Liebig-Weg 3, 37077 Göttingen, Germany
⁵ Skolkovo Institute of Science and Technology, Bolshoy Boulevard 30, bld. 1, Moscow 121205, Russia
⁶ NorthWest Research Associates, 3380 Mitchell Ln, Boulder, CO 80301, USA

Received: 23 February 2021
Accepted: 28 April 2021

Abstract

Context. A precise detection of the coronal hole boundary is of primary interest for a better understanding of the physics of coronal holes, their role in the solar cycle evolution, and space weather forecasting.

Aims. We develop a reliable, fully automatic method for the detection of coronal holes that provides consistent full-disk segmentation maps over the full solar cycle and can perform in real-time.

Methods. We use a convolutional neural network to identify the boundaries of coronal holes from the seven extreme ultraviolet (EUV) channels of the Atmospheric Imaging Assembly (AIA) and from the line-of-sight magnetograms provided by the Helioseismic and Magnetic Imager (HMI) on board the Solar Dynamics Observatory (SDO). For our primary model (Coronal Hole RecOgnition Neural Network Over multi-Spectral-data; CHRONNOS) we use a progressively growing network approach that allows for efficient training, provides detailed segmentation maps, and takes into account relations across the full solar disk.

Results. We provide a thorough evaluation for performance, reliability, and consistency by comparing the model results to an independent manually curated test set. Our model shows good agreement to the manual labels with an intersection-over-union (IoU) of 0.63. From the total of 261 coronal holes with an area > 1.5 × 10¹⁰ km² identified during the time-period from November 2010 to December 2016, 98.1% were correctly detected by our model. The evaluation over almost the full solar cycle no. 24 shows that our model provides reliable coronal hole detections independent of the level of solar activity. From a direct comparison over short timescales of days to weeks, we find that our model exceeds human performance in terms of consistency and reliability. In addition, we train our model to identify coronal holes from each channel separately and show that the neural network provides the best performance with the combined channel information, but that coronal hole segmentation maps can also be obtained from line-of-sight magnetograms alone.

Conclusions. The proposed neural network provides a reliable data set for the study of solar-cycle dependencies and coronal-hole parameters. Given the fast and robust coronal hole segmentation, the algorithm is also highly suitable for real-time space weather applications.