Solar activities have a great impact on our home planet and space environment. A huge amount of data has been collected by solar satellites, such as Solar Dynamic Observatory (SDO), to routinely record solar activities for better scientific understanding as well as for predicting their impact on space weather. One such feature or phenomenon on the Sun is the coronal hole and its evolution. Coronal holes, observed as a large-scale dark patch in EUV corona images, are known to be source regions of fast solar winds that shape the structure of the entire heliosphere and are responsible for geomagnetic storms at different levels.

AI/ML tools are increasingly used to aid research for a better understanding of solar activities. The training data for AI/ML models became a big gap in the process, as petabytes of solar images archived have not been systematically labeled yet. This system utilizes the NASA MADRAIT project results on auto labeling training datasets to automatically build training datasets for solar coronal hole study. A petabyte storage transferred from NASA Goddard to GMU is utilized to host the data for public access, and a similar instance will be hosted with NASA Goddard SDO archive.