Skip to Content

A Ground-Truth Video Dataset for the Evaluation of Vision-based Sense-and-Avoid systems

The development and evaluation of vision-based Sense-and-Avoid systems demand flight scenario images which are expensive and risky to obtain. Nowadays Augmented Reality techniques allow the compositing of real flight scenario images with 3D aircraft models to produce useful realistic images for system development and benchmarking purposes at a much lower cost and risk.

With the techniques we are presenting, first 3D aircraft models are positioned in a simulated 3D flight scene with controlled illumination and rendering parameters. Realistic simulated images are then obtained using an image processing algorithm which fuses the images obtained from the 3D scene with images from real UAV flights taking into account on board camera vibrations. The annotation of these images is a straightforward process since the pose of the simulated intruder aircraft is known. These ground truth annotations allow to develop and quantitatively evaluate aircraft detection and tracking algorithms.