CEAS EuroGNC 2026
|
Performance Assessment of AI-driven relative pose estimation algorithms: YOLO vs CenterPose
|
| Luis Rueda |
PhD Candidate, Universidad Rey Juan Carlos, Madrid, Spain. | | Hodei Urrutxua |
Associate Professor, Universidad Rey Juan Carlos, Madrid, Spain. | | Xin Chen |
Assistant Professor, Universidad Rey Juan Carlos, Madrid, Spain. | | Miguel Leiva |
PhD Candidate, Universidad Politécnica de Madrid, Madrid, Spain. | | Manuel Sanjurjo |
Full Professor, Universidad Carlos III de Madrid, Madrid, Spain. |
|
| Abstract:
The growing demand for autonomous rendezvous, inspection, and Active Debris Removal (ADR) missions calls for reliable vision-based relative navigation under challenging orbital conditions, where classical feature-based pipelines often fail. Deep learning provides a powerful alternative, as modern CNN/Transformer backbones can extract robust, illumination-tolerant features even under texture poverty, occlusions, or degenerate views. This work benchmarks three representative AI-based 6 Degrees of Freedom (6-DoF) pose estimation approaches for non-cooperative spacecraft: YOLOv8+SQPnP, the new YOLOv11+SQPnP, and NVIDIA CenterPose. A custom Blender-generated dataset of the Deimos-1 satellite comprising 16,200 photorealistic grayscale images with systematically varied viewing and illumination geometries was used for training and evaluation. Results in controlled fly-around scenarios show that YOLOv11+SQPnP achieves the best overall balance between geometric accuracy and temporal stability, with translation errors around 45-60 mm and orientation errors near 2-3 deg across most viewpoints. CenterPose DLA-34 remains the most robust under adverse illumination or self-shadowing, consistently maintaining low variance and smooth trajectories (translation errors ~52-55 mm, rotation errors ~3.5 deg). YOLOv8+SQPnP provides the sharpest geometric fits (3DIoU up to 0.92) but suffers from high sensitivity to viewpoint and lighting. Within this controlled synthetic benchmark, these findings indicate that compact detector-plus-PnP pipelines can offer a lower-complexity alternative to heavier integrated backbones and are promising candidates for future onboard assessment, although embedded suitability was not benchmarked here.
|
| Keywords: Relative pose estimation; Deep learning; Optical navigation; YOLO; CenterPose; Active Debris Removal; Proximity operations |
 View PDF CEAS-GNC-2026-075 doi: 10.82124/CEAS-GNC-2026-075 |
| Luis Rueda, Hodei Urrutxua, Xin Chen, Miguel Leiva, Manuel Sanjurjo: Performance Assessment of AI-driven relative pose estimation algorithms: YOLO vs CenterPose. Proceedings of the 2026 CEAS EuroGNC conference. Madrid, Spain. May 2026. doi: 10.82124/CEAS-GNC-2026-075.
|
| BibTeX entry (UTF-8):
@Incollection{CEAS-GNC-2026-075,
author = {Rueda, Luis and Urrutxua, Hodei and Chen, Xin and Leiva, Miguel and Sanjurjo, Manuel},
title = {Performance Assessment of AI-driven relative pose estimation algorithms: YOLO vs CenterPose},
booktitle = {Proceedings of the 2026 {CEAS EuroGNC} conference},
address = {Madrid, Spain},
month = may,
year = {2026},
doi = {10.82124/CEAS-GNC-2026-075}
}
|