Machine Learning for Integrating Ground-Based and Satellite Astronomical Data to Address Data Gaps Alfiah Rizky Diana Putri
Department of Atmospheric and Planetary Sciences, Faculty of Science, Institut Teknologi Sumatera, South
Abstract
^Ground-based astronomical data, such as astroclimate and light pollution measurements, are often used as ground truth. However, in equatorial regions, these data are frequently limited due to weather and cloud cover. This study explores the integration of satellite data from VIIRS-DNB, Luojia-1, and ECMWF ERA5 as training inputs for machine learning models, with in-situ SQM and weather sensor data as targets, to effectively bridge these data gaps. In this effort, better resolution in-situ data are made based on the relationship between the training and targets.
