Automatic Identification of Solar Active Areas

Active areas identified, and analysed according to their "texture" as an attempt at classification.

Monitoring and predicting space weather is heavily reliant on accurately identifying active areas on the sun and classifying them according to likelihood of flaring or mass ejection activity. Armed with a vast array of image processing and machine learning techniques, we are spoiled for choice when creating automatic identifiers of active areas in magnetograms. Classification of active areas is a somewhat more difficult task to automate. This project is co-supervised with collaborators at NWU Computer Science Department.