The presentation wil introduce an open source dataset tailored specifically for the aerospace industry, focusing on the detection of runways. Comprising 17,000 images, each equipped with relevant metadata, this dataset serves as a valuable resource for training neural networks to identify runways accurately. The creation process adheres to stringent Operational Domain Definition (ODD) requirements, incorporating images sourced from Google Earth and real-life landing videos on YouTube.

In addition to the comprehensive image collection, we present scripts that empower the community to expand and diversify the dataset further. By fostering collaboration, this dataset serves as a catalyst for pioneering research in object detection, instilling robustness, explainability, and confidence in artificial intelligence applications developed for aeronautical contexts.