Recognition Method for Mount Wilson Magnetic Type of Sunspots Based on Deep Learning

Sunspots are the regions with stronger magnetic field in the solar photosphere and most of solar eruptions occur in complex sunspot groups. Mount Wilson magnetic classification is one of the most popular sunspots classification methods, which is of great significance to the study of solar eruptions. In recent years, with the rapid development of China’s space industry, space physics research has entered the era of big data. Deep learning methods for processing space science data are springing up. In this study, based on the SDO/HMI SHARP continuum and magnetogram data during 2010－2017, we propose to apply deep learning for the image recognition of Mount Wilson magnetic type of sunspots. The results show that Xception has a productive performance in the identification of the sunspots magnetic types in solar active regions. The F1 score of sunspots group of α exceeds 96%, that of β is more than 93%, and that of other types is more than 84%.