PROCESSING OF METEOROLOGICAL IMAGES TO PREDICT FLIGHT HOLDING MANEUVERS

This work proposes the improvement of an existing model that predicts flight holding maneuvers based on meteorological data and satellite images. A holding maneuver is an airborne delay procedure that aircraft follow when they cannot land immediately, typically due to traffic congestion or adverse weather conditions. Predicting these maneuvers is important for optimizing air traffic management, minimizing delays, and ensuring safety. The original model crops a large static region encompassing all of Brazil, while this research introduces a localized approach that crops smaller, more relevant areas around the destination airport. Specifically, this study tests cropping images with areas of 500 km², 1000 km², and 2000 km² to enhance prediction accuracy, ultimately aiding in better flight management and reducing delays caused by adverse weather conditions.