Hao Liu
1,2,
Yan Xu
1,3,4,
Jiasheng Wang
1,3,4,
Ju Jing
1,3,4,
Chang Liu
1,3,4,
Jason T. L. Wang
1,2, and Haimin Wang
1,3,4
1. Institute for Space Weather Sciences, New Jersey Institute of
Technology
2. Department of Computer Science, New Jersey Institute of
Technology
3. Big Bear Solar Observatory, New Jersey Institute of Technology
4. Center for Solar-Terrestrial Research, New Jersey Institute of
Technology
Abstract
We propose a new machine learning approach to Stokes inversion based on a convolutional neural network (CNN) and the
Milne-Eddington (ME) method.
The Stokes measurements used in this study were taken by the Near InfraRed Imaging Spectropolarimeter (NIRIS)
on the 1.6 m Goode Solar Telescope (GST) at the Big Bear Solar Observatory.
By learning the latent patterns in the training data prepared by the physics-based ME tool,
the proposed CNN method is able to infer vector magnetic fields from the Stokes profiles of GST/NIRIS.
Experimental results show that
our CNN method produces smoother and cleaner magnetic maps than the widely used ME method.
Furthermore, the CNN method is 4~6 times faster than the ME method,
and is able to produce vector magnetic fields in near real-time,
which is essential to space weather forecasting.
Specifically, it takes ~50 seconds for the CNN method
to process an image of 720×720 pixels
comprising Stokes profiles of GST/NIRIS.
Finally, the CNN-inferred results are highly correlated to
the ME-calculated results and are closer to the ME's results
with the Pearson product-moment correlation coefficient (PPMCC) being closer to 1 on average
than those from other machine learning algorithms
such as multiple support vector regression and multilayer perceptrons (MLP).
In particular, the CNN method outperforms the current best machine learning method (MLP) by 2.6% on average
in PPMCC according to our experimental study.
Thus, the proposed physics-assisted deep learning-based CNN tool
can be considered as an alternative, efficient method
for Stokes inversion for high resolution polarimetric observations obtained by GST/NIRIS.
Source Code
»
Click
here
to download
the CNN tool (46MB) described in the paper.
Datasets
»
Click
here
to download
the training set named trainset.zip (in the
FITS format, 574MB) containing (i) training images
from AR 12371 collected on 2015 June 22, and (ii)
ME-calculated results used as labels for training.
»
Click
here
to download
the training set named traincsv.zip (in the
CSV format, 287MB) containing one million labeled
data samples converted from the
above FITS files where the data samples
are used to train the CNN tool.
»
Click
here
to download
the first test set named
testset1.zip (275MB) containing the
720×720 test image (in the FITS format) from AR 12371
collected on 2015 June 25 20:00:00 UT
and the CSV file converted from the FITS file.
»
Click
here
to download
the second test set named
testset2.zip (232MB) containing the
720×720 test image (in the FITS format) from
AR 12665 collected on 2017 July 13 18:35:00 UT
and the CSV file converted from the FITS file.
»
Click
here
to download
the third test set named
testset3.zip (244MB) containing the
720×720 test image (in the FITS format) from
AR 12673 collected on 2017 September 6 19:18:00 UT
and the CSV file converted from the FITS file.
»
Click
here
to download the CNN-inferred results
(in the CSV format and PNG format, 11MB)
obtained by applying the trained CNN model
to the test image from AR 12371 collected on 2015 June 25 20:00:00 UT.
»
Click
here
to download the CNN-inferred results
(in the CSV format and PNG format, 11MB)
obtained by applying the trained CNN model
to the test image from
AR 12665 collected on 2017 July 13 18:35:00 UT.
»
Click
here
to download the CNN-inferred results
(in the CSV format and PNG format, 11MB)
obtained by applying the trained CNN model
to the test image from
AR 12673 collected on 2017 September 6 19:18:00 UT.
Reference
Inferring Vector Magnetic Fields from Stokes Profiles of GST/NIRIS
Using a Convolutional Neural Network.
Liu, H., Xu, Y., Wang, J., Jing, J., Liu, C., Wang, J. T. L., Wang, H.,
ApJ., 894:70, 2020
[
GitHub]
[
MyGitHub]