Published March 23, 2022
| Version 1
Dataset
Open
Multi-Spectral Solar Images for Active Region Detection and Segmentation
- 1. Swansea University
- 2. Université de Toulon
- 3. Observatoire de Paris
Description
A dataset of solar active regions annotations, including object detection and object segmentation labels (i.e., bounding boxes and masks).
Data consists of both, ground-based (sourced from Observatoire de Paris, Meudon Spectroheliograph images) and space-based NASA SOHO EIT and MDI sensors (UAD).
We split these images into two groups, UAD (upper atmosphere data) and LAD (lower atmosphere data) according to their association with altitudes in the solar atmosphere, reflecting the proximity in sensed information in constituting spectral lines.
It is good to note that, segmentation masks are only available for UAD.
Please cite https://doi.org/10.1007/s00138-021-01261-y
Files
MLMT-CNN_Solar-Active-Regions_Bounding-Box_and_Segmentation_Annotations_for_Deep-Learning_Application.zip
Files
(6.6 GB)
| Name | Size | Download all |
|---|---|---|
|
md5:f5bba8bf0b08c2d8ffaae22ae8683810
|
6.6 GB | Preview Download |