GAP-GDRNet: Geometry-Aware Monocular Visual Pose Sensing on a Single-Target Synthetic Spacecraft Dataset

2026-07-02

AI Summary

This work addresses the challenge of monocular 6D pose estimation for non-cooperative spacecraft under adverse conditions such as low texture, illumination variations, and partial occlusions. To this end, the authors propose a geometry-aware attention-enhanced framework that integrates an Attention-based Feature Refinement (AFR) module with a Patch-wise Geometric Self-Attention (PGSA) module. Within a direct regression paradigm, the method fuses sparse geometric cues and structural priors, further refining pose accuracy through geometric self-attention in a Patch-PnP stage. A synthetic dataset generated with Blender provides mask images, coordinate maps, and pose labels for supervised training. Experimental results demonstrate that the proposed approach significantly outperforms existing methods on a single-object synthetic spacecraft dataset, exhibiting superior robustness and precision particularly in low-texture and partially occluded scenarios.

Abstract

Monocular relative pose sensing is a central perception problem in non-cooperative rendezvous and on-orbit servicing. In spacecraft images, however, weak surface texture, thin appendages, illumination changes, and partial occlusion often leave only sparse and unstable geometric evidence. This article presents GAP-GDRNet, a geometry-aware attention-enhanced framework for monocular RGB-based 6D pose sensing. The method follows the geometry-guided direct regression paradigm of GDR-Net and modifies two points in the pipeline: an attention-based feature refinement (AFR) module is placed before dense geometric prediction, and a patch-level geometric self-attention (PGSA) module is inserted into Patch-PnP. AFR reinforces global spacecraft structure together with local weak-texture cues; PGSA then relates downsampled geometric patches before final pose regression. A Blender-based annotation process supplies target masks, visible-region masks, dense model-coordinate maps, camera intrinsics, and 6D pose labels for supervised training.

Problem

• monocular pose estimation
• non-cooperative rendezvous
• geometric evidence
• spacecraft
• 6D pose sensing

Innovation

• geometry-aware
• attention mechanism
• monocular pose estimation
• 6D pose
• spacecraft rendezvous