Foreign Object Debris detection at aerodromes using YOLOv5
Jakub Suder, Tomasz Marciniak · 2024
The latest EASA recommendations from 2024 indicate the possibility of using machine learning techniques in aerodrome monitoring. The aim of the work was to analyze solutions and prepare FOD (Foreign Object Debris) object detection software on aerodromes. In order to implement the issue, a dataset of photos and video recordings for testing algorithms was developed. The dataset consists of 1480 photos showing FOD at aerodromes and photos of the runway itself, according to EASA and FAA recommendations. FOD detection was implemented using a classical, k-means method and effective YOLOv5.