This repository contains the source code to reproduce coronal holes segmentation in synoptic maps following the paper Machine-learning approach to identification of coronal holes in solar disk images and synoptic maps.
Dataset of SDO/AIA 193 Angstrom and CHs synoptic maps is available online at https://sun.njit.edu/coronal_holes/. Feel free to download and explore it in your research.
Clone the repository with its submodules
git clone --recursive https://github.com/observethesun/synoptic_maps.git
The code is based on helio framework. See the API documentation to learn more about its features.
To start building synoptic maps you will need a dataset of solar disk images in JPEG or FITS formats. Note that for JPEG images there will be required additional information on solar disk location. Follow the steps in the notebook Make_synoptic_maps to complete the process.
For CHs segmentation we apply a neural network model trained on solar disk images. Visit the repository coronal_holes to prepare the model. It is recommended to use the same source of solar disk images for model training and synoptic maps construction. Follow the notebook CHs_segmentation to apply the model for synoptic maps.
Illarionov E., Kosovichev А., Tlatov A. 2020, arXiv:2006.08529