Automatic Detection of Sunspots on Full-disk Solar Images Using the Simulated Annealing Genetic Method

Abstract

Sunspots with strong magnetic fields are the most important manifestations of solar activity, appearing as dark features in the photosphere observed in continuum images. We proposed an artificial intelligence technology called the simulated annealing genetic (SAG) method, which combined the genetic algorithm and simulated annealing algorithm to self-adaptively derive dual thresholds for detecting the umbra and penumbra of sunspots simultaneously. Full-disk continuum intensity images obtained from Solar Dynamics Observatory/Helioseismic Magnetic Imager (HMI) at a cadence of four hours from 2010 May to 2016 December were used. The detection results showed that the dual thresholds derived by the SAG method have outstanding performance in segmenting the umbra and penumbra from the photosphere with a satisfactory robustness efficiently. The boundaries of the umbra and penumbra were finely delineated, even for sunspots at the extreme solar limb. The total sunspot areas, umbral areas, and penumbral areas match very well with the data reported from HMI Debrecen Data (HMIDD), with the correlation coefficients reaching 0.99, 0.99, and 0.95, respectively. The mean ratios of umbra to sunspot areas per year ranged from 0.159 to 0.233. The ratios decreased with an increase in solar activity, which implies that the ratio was related to the solar activity level.