MGAW: An Effective Method for Geo-localization in Adverse Weather

Aerial-view geo-localization involves determining an unknown location by comparing drone-captured images with geo-tagged satellite images. Traditional deep learning-based methods falter under adverse weather conditions like rain and fog due to insufficient adaptation to the domain shift between training scenarios and diverse test environments. To overcome this limitation, we introduce a new data-driven framework designed to adapt to various domain transformations by effectively extracting features from complex environments. Our solution enhances model performance through adaptive data augmentation, achieving third place in the Multimedia Drone Satellite Matching Challenge across multiple environments in the University-160k-WX dataset.