Automated Detection, Characterization, and Tracking of Sunspots from SoHO/MDI Continuum Images.

We describe a procedure for automated detection of sunspots from SoHO/MDI full-disk continuum images. The MDI Level-1.8 continuum images were first corrected for the limb darkening and stray light, and then were flat-fielded. Sunspots were extracted using a newly developed automated sunspot detection procedure, which is based on the level set, namely the selective binary and Gaussian function regularized level set (SBGFRLS) method (Zhang et al., Image Vis. Comput.28, 668, 2010 ). In this method we initialize a two-dimensional level-set function and evolve it using a signed pressure force (SPF) function. For sunspot detection, the level-set function was defined twice, first for umbra and then for penumbra extraction. Using this procedure, along with the characterization of detected sunspots we have also generated tracking reports of all sunspots in a fully unsupervised manner. [ABSTRACT FROM AUTHOR]