CAD-Free Learning of Spacecraft Pose Estimators via NeRF-Based Augmentations

2026-05-19

AI Summary

This work addresses the limited generalization of traditional spacecraft pose estimation methods, which rely heavily on CAD-rendered images and struggle in on-orbit scenarios with unreliable geometric priors or significant discrepancies in real-world lighting and material appearance. The study introduces neural radiance fields (NeRF) to this task for the first time, enabling reconstruction of a target’s 3D representation from only 25 to 400 real images. By leveraging geometrically consistent view synthesis and appearance augmentation, the method generates large-scale training data without requiring any CAD models, thereby training a high-precision pose estimator. This approach substantially improves robustness under complex illumination conditions and, when combined with CAD data, further enhances out-of-domain generalization performance.

Abstract

Spacecraft pose estimation networks require tens of thousands of CAD-rendered images to be trained. This reliance on synthetic CAD data (i) limits applicability to targets with reliable geometry prior, excluding uncooperative or poorly documented spacecraft, and (ii) causes poor generalization to real on-orbit conditions due to unrealistic illumination and material appearance. This paper introduces a NeRF-based image augmentation method that enables the learning of spacecraft pose estimators from only a few tens to a few hundreds of images. The method learns a Neural Radiance Field of the target and generates a large, diverse dataset through geometrically-consistent viewpoint and appearance augmentation. This augmented dataset enables the training of accurate target-specific pose estimators without requiring a CAD model or large synthetic datasets. Experiments show that our approach supports the training of accurate pose estimators from only 25 to 400 realistic images, even under severe illumination variations. When applied on large CAD-based synthetic datasets, the NeRF-based augmentation also enhances out-of-domain generalization, yielding improved robustness to real on-orbit conditions.

Problem

• spacecraft pose estimation
• CAD-free
• domain generalization
• realistic image augmentation
• on-orbit conditions

Innovation

• NeRF-based augmentation
• CAD-free pose estimation
• spacecraft pose estimation
• geometrically-consistent synthesis
• realistic image augmentation