Segmentation of coronal holes in solar disc images with a convolutional neural network

Illarionov, Egor; Tlatov, A. G.

doi:10.1093/mnras/sty2628

Cited by 40 publications

(35 citation statements)

References 31 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…U-nets are widely used for image segmentation, e.g., the segmentation of coronal holes in solar data (Illarionov and Tlatov, 2018). This neural network inherits its name from the shape of its architecture which features a contracting branch, a bottleneck and an expansive branch (Ronneberger et al, 2015).…”

Section: Methodsmentioning

confidence: 99%

Inferring Plasma Flows at Granular and Supergranular Scales With a New Architecture for the DeepVel Neural Network

Tremblay

Attié

2020

Front. Astron. Space Sci.

View full text Add to dashboard Cite

The wealth of observational data available has been instrumental in investigating physical features relevant to solar granulation, supergranulation and Active Regions. Meanwhile, numerical models have attempted to bridge the gap between the physics of the solar interior and such observations. However, there are relevant physical quantities that can be modeled but that cannot be directly measured and must be inferred. For example, direct measurements of plasma motions at the photosphere are limited to the line-of-sight component. Methods have consequently been developed to infer the transverse plasma motions from continuum images in the case of the Quiet Sun and magnetograms in the case of Active Regions. Correlation-based tracking methods calculate the optical flows by correlating series of images locally while other methods like "Coherent Structure Tracking" or "Balltracking" exploit the coherency of photospheric granules to track them and use the group motions of the granules as a proxy of the average plasma flows advecting them. Recently, neural network computing has been used in conjunction with numerical models of the Sun to be able to recover the full velocity vector in photospheric plasma from continuum images. We experiment with a new architecture for the DeepVel neural network which takes inspiration from the U-Net architecture. Simulation data of the Quiet Sun and Active Regions are then used to evaluate the response at granular and supergranular scales of the aforementioned method.

show abstract

Section: Methodsmentioning

confidence: 99%

Inferring Plasma Flows at Granular and Supergranular Scales With a New Architecture for the DeepVel Neural Network

Tremblay

Attié

2020

Front. Astron. Space Sci.

View full text Add to dashboard Cite

show abstract

“…Studies comparing the two different conceptual approaches have shown significant differences in the size, location, shape and occurrence of the dark and/or open structures defined as CHs (e.g., Lowder et al, 2014;Lowder, Qiu, and Leamon, 2017;Linker et al, 2017;Wallace et al, 2019;Huang, Lin, and Lee, 2019;Asvestari et al, 2019). Additionally new approaches like machine learning/neural networks (e.g., Illarionov and Tlatov, 2018) and extraction methods based on plasma properties (differential emission measure; Raymond and Doyle, 1981;Hahn, Landi, and Savin, 2011) are the topic of current research.…”

Section: Introductionmentioning

confidence: 99%

Statistical Analysis and Catalog of Non-polar Coronal Holes Covering the SDO-Era Using CATCH

et al. 2019

View full text Add to dashboard Cite

Coronal holes are usually defined as dark structures seen in the extreme ultraviolet and X-ray spectrum which are generally associated with open magnetic fields. Deriving reliably the coronal hole boundary is of high interest, as its area, underlying magnetic field, and other properties give important hints as regards high speed solar wind acceleration processes and compression regions arriving at Earth. In this study we present a new threshold-based extraction method, which incorporates the intensity gradient along the coronal hole boundary, which is implemented as a user-friendly SSW-IDL GUI. The Collection of Analysis Tools for Coronal Holes (CATCH) enables the user to download data, perform guided coronal hole extraction and analyze the underlying photospheric magnetic field. We use CATCH to analyze non-polar coronal holes during the SDO-era, based on 193 Å filtergrams taken by the Atmospheric Imaging Assembly (AIA) and magnetograms taken by the Heliospheric and Magnetic Imager (HMI), both on board the Solar Dynamics Observatory (SDO). Between 2010 and 2019 we investigate 707 coronal holes that are located close to the central meridian. We find coronal holes distributed across latitudes of about ± 60 • , for which we derive sizes between 1.6 × 10 9 and 1.8 × 10 11 km 2. The absolute value of the mean signed magnetic field strength tends towards an average of 2.9 ± 1.9 G. As far as the abundance and size of coronal holes is concerned, we find no distinct trend towards the northern or southern hemisphere. We find that variations in local and global conditions may significantly change Electronic supplementary material The online version of this article (

show abstract

“…A multi-wavelength approach was developed by Garton et al (2018) in the form of the multi-thermal emission recognition algorithm CHIMERA. Recently, with the dawn of machine learning, new methods, utilizing the increased computational performance have also emerged to provide an additional tool to identify and extract coronal holes (e.g., Illarionov and Tlatov 2018).…”

Section: Stream Interaction Regions/co-rotating Interaction Regions (mentioning

confidence: 99%

A review of the SCOSTEP’s 5-year scientific program VarSITI—Variability of the Sun and Its Terrestrial Impact

Shiokawa

Georgieva

2021

Prog Earth Planet Sci

View full text Add to dashboard Cite

The Sun is a variable active-dynamo star, emitting radiation in all wavelengths and solar-wind plasma to the interplanetary space. The Earth is immersed in this radiation and solar wind, showing various responses in geospace and atmosphere. This Sun–Earth connection variates in time scales from milli-seconds to millennia and beyond. The solar activity, which has a ~11-year periodicity, is gradually declining in recent three solar cycles, suggesting a possibility of a grand minimum in near future. VarSITI—variability of the Sun and its terrestrial impact—was the 5-year program of the scientific committee on solar-terrestrial physics (SCOSTEP) in 2014–2018, focusing on this variability of the Sun and its consequences on the Earth. This paper reviews some background of SCOSTEP and its past programs, achievements of the 5-year VarSITI program, and remaining outstanding questions after VarSITI.

show abstract

Segmentation of coronal holes in solar disc images with a convolutional neural network

Cited by 40 publications

References 31 publications

Inferring Plasma Flows at Granular and Supergranular Scales With a New Architecture for the DeepVel Neural Network

Inferring Plasma Flows at Granular and Supergranular Scales With a New Architecture for the DeepVel Neural Network

Statistical Analysis and Catalog of Non-polar Coronal Holes Covering the SDO-Era Using CATCH

A review of the SCOSTEP’s 5-year scientific program VarSITI—Variability of the Sun and Its Terrestrial Impact

Contact Info

Product

Resources

About