Automatic Sunspot Classification for Real-Time Forecasting of Solar Activities

Solar imaging is currently an active area of research. A fast hybrid system for the automated detection and classification of sunspot groups on MDI Continuum images using active regions data extracted from MDI Magnetogram images is presented in this paper. The system has three major stages: sunspots detection from MDI Continuum images, sunspots grouping and McIntosh classification of sunspot groups. Image processing and machine learning are integrated in all these stages.