SOLAR ACTIVE REGION MAGNETOGRAM IMAGE DATASET

本数据集名为SOLAR ACTIVE REGION MAGNETOGRAM IMAGE DATASET，由新墨西哥州立大学电气与计算机工程学院创建。数据集包含来自NASA太阳动力学观测站(SDO)的磁力图，这些图像量化了太阳磁场的强度。数据集整合了三个来源的数据，提供了太阳活动区域(磁通量大的区域)的SDO日震和磁成像仪(HMI)磁力图以及相应的耀斑活动标签。数据集大小为1,372,004张图像，主要用于太阳物理研究，如磁结构的图像分析、其随时间的演变及其与太阳耀斑的关系。数据集的创建过程涉及从NOAA空间天气预测中心(SWPC)的太阳区域摘要(SRS)提取数据，通过联合科学操作中心(JSOC)界面下载磁力图图像，并使用SWPC事件报告(ER)提取与耀斑活动相关的标签。数据集的应用领域包括自动太阳耀斑预测方法的研究，包括监督和无监督机器学习、二元和多类分类以及回归分析。

This dataset is named SOLAR ACTIVE REGION MAGNETOGRAM IMAGE DATASET, and was created by the College of Electrical and Computer Engineering, New Mexico State University. The dataset contains magnetograms from NASA's Solar Dynamics Observatory (SDO), which quantify the strength of the solar magnetic field. It integrates data from three sources, providing SDO Helioseismic and Magnetic Imager (HMI) magnetograms of solar active regions (flux-dense regions) alongside corresponding flare activity labels. Comprising 1,372,004 images in total, this dataset is primarily utilized for solar physics research, such as image analysis of magnetic structures, their temporal evolution, and their correlations with solar flares. The dataset creation process involved extracting data from the Solar Region Summaries (SRS) of the NOAA Space Weather Prediction Center (SWPC), downloading magnetogram images via the Joint Science Operations Center (JSOC) interface, and extracting flare activity-related labels using SWPC Event Reports (ER). Its application scope covers research on automated solar flare prediction methods, including supervised and unsupervised machine learning, binary and multi-class classification, and regression analysis.