Detecting coronal holes for solar activity modeling

Solar image analysis relies on the detection of coronal holes for predicting disruptions to earth´s magnetic field. This paper introduces a level-set method for detecting coronal holes based on the processing of extreme ultra-violet images (EUVI) and magnetic images. For validating the approach, two independent manual annotations were combined to produce a set of 46 consensus maps. Overall, the level-set method produces significant improvements over the currently used approach. Future work needs to focus on validating the approach on larger datasets, the integration of more imaging modalities, and an analysis of inter-rater and intra-rater variability.