URSO: Unreal Rendered Spacecrafts On-Orbit Datasets

Synthetic image datasets of Soyuz and Dragon spacecraft models orbiting the Earth, rendered using our simulator, to learn and evaluate pose estimation algorithms. For more information, please refer to our preprint: https://arxiv.org/abs/1907.04298

Each dataset contains:

• 5000 RGB images with a resolution of 1280x960 px
• Train/val/test set splits as a list: <set>_images.csv
• Pose ground truth as: <set>_poses_gt.csv with
  • Row format: <x y z qx qy qz qw>

The camera calibration matrix is:

\(K = \begin{bmatrix} 640.00 & 0 & 640 \\ 0 & 640.46 & 480\\ 0 & 0 & 1 \end{bmatrix}\)

And the horizontal and vertical field-of-view are respectively 90.0° and 73.7º.