Perpustakaan Digital ITB

Air Traffic Control (ATC) at airports is sometimes challenged by fog and rain, which impair the detection and tracking of aircraft, especially during landing. This study presents a deep learning approach employing vision-AI YOLO (You Only Look Once) to improve aircraft detection under low-visibility conditions. A dataset of 1,001 images was augmented with a simulated fog effect to mimic adverse weather conditions, and two distinct detection models were trained using this enhanced dataset. Experimental results indicate that the proposed models achieve high precision, recall, F1-score, and mAP, demonstrating robust performance in challenging visual environments. The findings suggest that incorporating fog-effect augmentation can significantly enhance aircraft detection and tracking in low-visibility conditions, with an mAP@0.50–0.95 score of 81.57%. This approach improves ATC decision-making, offering a scalable solution for aviation safety at airports.