Airport Drone Threat & Safety Dataset

Rogue drones near airports present a critical safety threat, yet capturing real training data in restricted airspace is nearly impossible. This synthetic dataset solves that challenge by generating photorealistic drone imagery against complex airport backgrounds including terminals, runways, hangars, and control towers. Drones appear as small objects occupying less than 5% of the frame, simulating real detection challenges. The dataset includes adverse weather conditions (fog, rain, overcast, nighttime) and varied drone types. Annotations in YOLO format include drone bounding boxes and optional bird-vs-drone classification labels for reducing false positives. The open-source sample contains 111 images; the full package offers 1,200 images with evaluation videos.