Automation of all-sky images classification

The Earth's ionosphere is frequently perturbed by solar, geomagnetic, and atmospheric events, which can impact critical technologies. To understand the sources, development, and consequences of ionospheric phenomena, extensive statistics of detected events are required. Currently, a vast repository of such events exists in the data archives of all-sky airglow imagers operated by Boston University. However, the challenge lies in automatically identifying relevant images. The application of convolutional neural networks offers an effective solution to bridge this gap. The poster will present insights derived from OI(630.0 nm) airglow measurements at the low-latitude El Leoncito Observatory (Argentina).