Jul 2026· IEEE International Workshop on Metrology for AeroSpace· pp. 70-75· 0 citations· 17 references
Abstract
Accurate monocular pose estimation of noncooperative spacecraft is critical for autonomous proximity operations such as on-orbit servicing and active debris removal. A major practical challenge is the domain gap between synthetic training imagery and real Hardware-In-the-Loop (HIL) sensor data. We present a pose estimation pipeline built on a self-supervised DINOv3 Vision Transformer backbone with a deconvolutional heatmap head that localizes spacecraft keypoints. To bridge the synthetic-to-real gap without target-domain annotations, the pipeline combines multi-scale structural similarity (MS-SSIM) supervision, feature-level domain generalization, input-level style randomization, and iterative self-training with pseudo-labels from unlabeled HIL images. On the SPEED+ benchmark, the pipeline achieves state-of-the-art accuracy on both HIL domains using a single model with no adversarial training, no external augmentation networks, and no target-domain annotations.
Abstract. This paper presents a scalable framework for real-time, 6-DOF pose estimation of uncooperative space objects using monocular cameras. We address dataset scarcity by generating mission-specific synthetic data augmented with style randomization to mitigate domain shift. A lightweight YOLOv11n-pose model extracts keypoints, which feed into a RANSAC-PnP solver for initial pose estimation, refined through an Extended Kalman Filter leveraging rigid body dynamics. Validation against synthetic sequences and laboratory experiments with a 1:1 replica demonstrates mean orientation errors of 3.9° and translation errors of 1.1%, matching state-of-the-art performance. The pipeline maintains robustness under challenging lighting conditions and domain shift, achieving ~50 fps on GPU and ~5 fps on CPU, enabling deployment on resource-constrained platforms for on-orbit servicing and debris removal missions.
Matteo Forasassi· Materials Research Proceedin...· 0 citations
This work presents the first complete system for automated six degrees of freedom (6DOF) satellite pose estimation from spatially resolved, ground-based, adaptive optics (AO)-corrected imagery, addressing a key challenge in Space Domain Awareness (SDA). The approach mitigates the need for human labeling by directly regressing satellite orientation and position from blurry, noisy, and deeply shadowed imagery. A multi-stage deep neural network pipeline localizes the satellite, predicts pose, and optionally applies temporal filtering. Networks are trained exclusively on fully synthetic imagery generated from a CAD model, yet generalize effectively to real data, bridging the Sim2Real domain gap. On 137 real, human-labeled test images of Seasat, the model achieved a mean rotation error of 5° and a mean image-plane translation error of 21 cm. Slant range error was quantitatively evaluated on synthetic data due to unknown real-sensor parameters. Qualitative evaluation of additional real Seasat imagery rated 177 of 199 predicted poses as “ground truth equivalent” or “high-confidence match,” with zero catastrophic failures. The system was extended to seven degrees of freedom (7DOF) for satellites with articulating components and demonstrated on real Hubble Space Telescope (HST) imagery, achieving 5.5° rotation error, 51 cm image-plane translation error, and 8° symmetry-adjusted solar array error on a 249-frame pass with causal temporal filtering. Across 586 real test images from Seasat and HST (captured over multiple decades under diverse conditions) the system consistently performed well. Full 6DOF performance was quantified on a high-fidelity wave optics (HFWO) synthetic test set of Seasat, where the model achieved 8.4° mean rotation error, 34 cm image-plane translation error, and 1.4% line-of-sight range error at r0=6 cm and 1031 km range. In a limited 200-image benchmark, the model demonstrated 48% lower mean rotation error than a single human labeler while operating ∼800× faster. It required <40 h and a single A100 GPU to generate data and train. The approach was also demonstrated for ARGOS, a smaller satellite with highly symmetric geometry. An exploratory General Image-Quality Equation-based image quality metric (AO-IQ) was introduced as an empirical correlate for pose accuracy. General-purpose models like GPT-4o and Depth Anything V2 failed across most SDA tasks, but rapid gains in vision-language models warrant continued monitoring. These results establish a new operational baseline for practical, real-time satellite pose estimation from AO SDA imagery.
Thomas J. Dickinson, Dawson Friesenhahn, Justin Fletcher et al.· Aerospace· 0 citations
Comprehensive comparisons with CNN-, Transformer- and prevailing Mamba-based approaches verify the advantages of SCI-Mamba in visual authenticity, color fidelity and inference speed, and provides a practical low-light enhancement solution for close-proximity non-cooperative space operations.
Yiyong Sun, Weihang Shan, Shijun Wei et al.· 0 citations
6-DoF pose estimation is a critical task in autonomous rendezvous and proximity operations. In the case of an unknown target, this task becomes challenging as it shall be paired with the reconstruction of the target shape model. In this article, we propose a novel framework for single-shot shape and pose estimation of unknown spacecraft objects. Given a single image, we first reconstruct a 3D shape model of the target, then estimate the relative six-degrees-of-freedom pose by learning dense 2D-3D correspondences. The image features are extracted using a frozen DINOv3 vision transformer, while the geometric features are computed from the reconstructed point cloud using a trainable dynamic graph convolutional neural network encoder. A dual-stream transformer matcher refines descriptors through alternating self- and cross-attention, producing soft correspondences that are passed to a Perspective-$n$-Point solver for pose recovery. We evaluate the method on the SPE3R dataset and consider FoundationPose as a representative baseline for current state-of-the-art capabilities. Results show reliable pose estimates achieving 0.157 degrees mean pointing error using only a single image and reconstructed geometry, demonstrating strong generalization to unseen spacecraft.
Josiane Uwumukiza, Jocelyn Zhao, Giovanni Lavezzi et al.· 0 citations
In the framework of space missions involving Close-Proximity Operations (CPO), accurate pose estimation between a chaser spacecraft and a non-cooperative resident space object remains a significant challenge. In such scenarios, pose determination must rely solely on Electro-Optical sensors such as cameras or LiDARs. Recently, Deep Learning (DL) approaches have gained significant attention for this task, demonstrating promising performance in terms of accuracy and runtime efficiency. This work presents a comparative analysis of both original and state-of-the-art point-based neural network architectures for LiDAR-based pose initialization of noncooperative spacecraft. The evaluated models, based on PointNet, differ in encoder adopted - PointNet (PN), a two-layer PointNet-based encoder (PN-PCN), and PointNet++ (PN++) - and in attitude parametrization, including quaternion, axisangle, 6D, and soft-label representations. The predicted pose is refined using the Iterative Closest Point (ICP) algorithm. Performance is assessed on synthetic scans of different targets generated by an in-house realistic LiDAR simulator. Results show that the PN-PCN encoder provides the best accuracyspeed trade-off, while soft-label parametrization yields the most accurate attitude estimation. With this configuration, subdegree-level attitude errors are achieved for Envisat and Aura, while degree-level attitude errors are obtained for Apollo, due to its more challenging symmetric geometry. Translational errors remain within a decimeter-level range across all targets.
Clemente Tecchia, A. Nocerino, G. Fasano et al.· IEEE International Workshop...· 0 citations
Monocular spacecraft 6D pose estimation remains difficult under weak texture, thin structures, illumination variation, and occlusion. This article presents GAP-GDRNet, a geometry-aware RGB framework built on GDR-Net for a single-target synthetic spacecraft benchmark. The method strengthens the geometry-guided regression pipeline at two points. First, AFR is placed before dense geometric prediction to combine global structural attention with local weak-texture enhancement. Second, PGSA is inserted into Patch-PnP to relate downsampled geometric regions before final pose regression. Dense supervision is obtained from a Blender-based rendering and annotation process that provides masks, model-coordinate maps, camera intrinsics, and 6D pose labels. On the self-built spacecraft dataset, GAP-GDRNet achieves a rotation error of 1.96{\deg}, a translation error of 0.0165 m,and 95.16% ADD@0.02 m, outperforming the reproduced GDR-Net baseline by 3.88 percentage points while running at 35.97 FPS. Tests on T-LESS and LM-O further show consistent gains over the reproduced baseline on textureless and occluded non-spacecraft objects.
Zongwu Xie, Yonglong Zhang, Yifan Yang et al.· 0 citations