pith. sign in

arxiv: 2605.18447 · v3 · pith:LIVWZLEVnew · submitted 2026-05-18 · 💻 cs.CV

NeRF-based Spacecraft Reconstruction from Monocular Imagery Under Illumination Variability and Pose Uncertainty

Antoine Legrand , Renaud Detry , Christophe De Vleeschouwer This is my paper

Pith reviewed 2026-05-21 07:56 UTC · model grok-4.3

classification 💻 cs.CV
keywords NeRFspacecraft reconstructionmonocular imageryillumination variabilitypose uncertainty3D reconstructionneural radiance fieldsin-orbit operations
0
0 comments X

The pith

Extending NeRF with per-image appearance embeddings and pose corrections enables robust 3D spacecraft reconstruction from monocular images despite illumination changes and pose inaccuracies.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper establishes that Neural Radiance Fields can be extended with two per-image learnable elements to handle the variable lighting and inaccurate pose labels common in monocular images of unknown spacecraft. An appearance embedding captures each image's specific illumination conditions while a pose correction refines the given label to improve alignment in 3D space. These additions are optimized jointly with the base model, adding little complexity yet yielding better consistency across the image collection. A sympathetic reader would care because such reconstructions support critical autonomous operations like debris removal and servicing where perfect data cannot be guaranteed.

Core claim

By introducing a learnable appearance embedding for each image to capture its unique illumination conditions and an image-specific pose correction term to refine noisy pose labels, the extended Neural Radiance Field achieves substantially better 3D consistency and reconstruction quality when trained on sets of monocular images with variable lighting and inaccurate poses. These parameters are learned jointly with the base model, adding only minimal complexity. Experiments on three image sets representative of in-orbit operations confirm the gains in robustness for offline reconstruction and suggest applicability to online settings.

What carries the argument

Per-image learnable appearance embedding and pose correction term, which adapt to illumination variability and refine pose estimates while jointly optimizing the radiance field for improved 3D consistency across images.

If this is right

  • The method supports reliable offline 3D modeling of uncooperative spacecraft for active debris removal and on-orbit servicing missions.
  • Validation on three representative in-orbit image sets demonstrates effectiveness under real illumination variability and pose uncertainty.
  • The joint optimization with minimal added parameters highlights suitability for handling imperfect real-world monocular data in radiance field applications.
  • The design points toward feasibility for extending the approach to online reconstruction scenarios during operations.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The minimal-complexity design could allow straightforward integration into existing NeRF pipelines for other monocular reconstruction tasks with lighting or pose noise.
  • If the pose corrections remain stable across larger sets, the method might reduce dependence on high-precision external pose sensors in future space missions.
  • The open mention of online reconstruction suggests testing sequential updates to the embeddings as new images arrive during an approach.

Load-bearing premise

Jointly optimizing the added per-image appearance embeddings and pose corrections with the base NeRF will yield substantial robustness gains on representative in-orbit image sets without introducing overfitting or optimization instabilities.

What would settle it

A side-by-side reconstruction on one of the three in-orbit image sets with and without the per-image terms, checking whether 3D consistency metrics and visual quality improve markedly only in the version that includes the learnable appearance embeddings and pose corrections.

read the original abstract

Autonomous rendezvous and proximity operations around uncooperative, unknown spacecraft are critical for active debris removal and on-orbit servicing missions. A key component of such operations is the offline reconstruction of a 3D model of the target from a set of 2D images. This task is challenging due to two main factors. First, in-orbit illumination conditions exhibit considerable variability, and change rapidly over time. Second, the inaccuracy of pose information in the images, results in 3D reconstruction uncertainty. To overcome these challenges, we propose to extend Neural Radiance Fields with per-image degrees of freedom: a learnable appearance embedding that captures the illumination conditions specific to each image, and an image-specific pose correction term that refines its noisy pose label to increase 3D consistency across images. These parameters add minimal complexity, as they are learned jointly with the NeRF, yet they substantially improve robustness to illumination variability and pose inaccuracies. We validate our approach on three image sets representative of in-orbit operations, demonstrating its effectiveness for offline reconstruction and highlighting its suitability for online reconstruction, an open problem in the field.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 1 minor

Summary. The paper claims to extend Neural Radiance Fields for offline 3D spacecraft reconstruction from monocular imagery by adding two per-image degrees of freedom learned jointly with the base NeRF: a learnable appearance embedding that captures image-specific illumination conditions, and an image-specific pose correction term that refines noisy pose labels to improve cross-image 3D consistency. It asserts that these additions add minimal complexity yet substantially improve robustness to illumination variability and pose inaccuracies, with validation on three representative in-orbit image sets.

Significance. If the central claims hold with proper controls, the work would be significant for space robotics applications such as active debris removal and on-orbit servicing, where variable illumination and inaccurate poses are common. The joint-optimization strategy for appearance and pose parameters is a practical strength that could support both offline and online reconstruction pipelines.

major comments (2)
  1. [Method] Method section (pose correction parameterization): the image-specific pose correction term is introduced to increase 3D consistency, yet the manuscript provides no explicit bounds, small-angle assumptions, or regularization loss penalizing large corrections. This is load-bearing for the central claim because, without such constraints, joint optimization with the appearance embedding and photometric loss can reduce rendering error by shifting poses to absorb illumination mismatches rather than enforcing multi-view geometry.
  2. [Experiments] Experiments section (quantitative validation): the abstract states effectiveness on three image sets but the manuscript description lacks reported quantitative metrics, baseline comparisons, or error analysis (e.g., pose error reduction or reconstruction quality scores). This prevents verification that the added terms deliver the claimed robustness gains rather than fitting noise.
minor comments (1)
  1. [Abstract] Abstract: the claim that the added parameters 'add minimal complexity' would benefit from a brief statement of the exact increase in parameter count or training time relative to standard NeRF.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the constructive feedback and for recognizing the potential significance of this work for space robotics applications. We address each major comment below and will revise the manuscript accordingly to strengthen the presentation of the method and results.

read point-by-point responses

  1. Referee: [Method] Method section (pose correction parameterization): the image-specific pose correction term is introduced to increase 3D consistency, yet the manuscript provides no explicit bounds, small-angle assumptions, or regularization loss penalizing large corrections. This is load-bearing for the central claim because, without such constraints, joint optimization with the appearance embedding and photometric loss can reduce rendering error by shifting poses to absorb illumination mismatches rather than enforcing multi-view geometry.

    Authors: We agree that explicit constraints on the pose correction are important to ensure it primarily enforces multi-view geometric consistency rather than absorbing illumination effects. In the revised manuscript we will add a regularization term to the total loss that penalizes large deviations from the initial pose estimates. We will also explicitly adopt and state a small-angle parameterization for the corrections, which is justified by the typical magnitude of pose uncertainty in the in-orbit datasets considered. These changes will be detailed in the Method section. revision: yes

  2. Referee: [Experiments] Experiments section (quantitative validation): the abstract states effectiveness on three image sets but the manuscript description lacks reported quantitative metrics, baseline comparisons, or error analysis (e.g., pose error reduction or reconstruction quality scores). This prevents verification that the added terms deliver the claimed robustness gains rather than fitting noise.

    Authors: We acknowledge the need for clearer quantitative evidence. The revised Experiments section will report reconstruction quality metrics (PSNR and SSIM) on the three image sets, direct comparisons against a baseline NeRF without the appearance embeddings or pose corrections, and quantitative analysis of the reduction in pose error achieved by the joint optimization. These additions will allow readers to verify the claimed robustness improvements. revision: yes

Circularity Check

0 steps flagged

No circularity: joint optimization of added parameters is independent of claimed robustness gains

full rationale

The paper's core proposal is to augment a standard NeRF with two additional per-image learnable variables (appearance embedding and pose correction) that are optimized jointly with the radiance field. No equations or derivations in the provided text reduce the target 3D consistency or robustness metric to these parameters by definition, nor do they invoke self-citations to establish uniqueness or forbid alternatives. The improvement claim rests on empirical validation across three representative image sets rather than on any fitted-input-renamed-as-prediction or self-definitional construction. This is a conventional extension pattern in neural rendering literature and remains self-contained against external benchmarks.

Axiom & Free-Parameter Ledger

2 free parameters · 2 axioms · 0 invented entities

The approach rests on standard NeRF scene representation plus domain assumptions about static targets and the utility of per-image corrections; two sets of learnable parameters are introduced as free degrees of freedom.

free parameters (2)
  • learnable appearance embedding
    Per-image vector capturing illumination conditions, optimized jointly with the radiance field.
  • image-specific pose correction term
    Additive refinement to noisy pose labels, learned to improve multi-view consistency.
axioms (2)
  • standard math A scene can be represented by a continuous radiance and density field queried along rays
    Core assumption inherited from the original NeRF formulation.
  • domain assumption The target spacecraft remains static across the captured image sequence
    Required for consistent 3D reconstruction from multiple views.

pith-pipeline@v0.9.0 · 5731 in / 1313 out tokens · 56478 ms · 2026-05-21T07:56:15.260882+00:00 · methodology

Review history (2 revisions) →

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Lean theorems connected to this paper

Citations machine-checked in the Pith Canon. Every link opens the source theorem in the public Lean library.

What do these tags mean?
matches
The paper's claim is directly supported by a theorem in the formal canon.
supports
The theorem supports part of the paper's argument, but the paper may add assumptions or extra steps.
extends
The paper goes beyond the formal theorem; the theorem is a base layer rather than the whole result.
uses
The paper appears to rely on the theorem as machinery.
contradicts
The paper's claim conflicts with a theorem or certificate in the canon.
unclear
Pith found a possible connection, but the passage is too broad, indirect, or ambiguous to say the theorem truly supports the claim.

Forward citations

Cited by 2 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

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

    cs.CV 2026-05 unverdicted novelty 6.0

    NeRF augmentation trains accurate spacecraft pose estimators from 25-400 real images without CAD models or large synthetic datasets.

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

    cs.CV 2026-05 unverdicted novelty 6.0

    NeRF-based image augmentation enables accurate target-specific spacecraft pose estimators to be trained from only 25-400 real images without CAD models or large synthetic datasets.

Reference graph

Works this paper leans on

61 extracted references · 61 canonical work pages · cited by 1 Pith paper

  1. [1]

    History of space shuttle rendezvous and proximity operations,

    Goodman, J. L., “History of space shuttle rendezvous and proximity operations,”Journal of Spacecraft and Rockets, Vol. 43, No. 5, 2006, pp. 944–959

    work page 2006

  2. [2]

    Optimal rendezvous trajectories of a controlled spacecraft and a tumbling object,

    Boyarko, G., Yakimenko, O., and Romano, M., “Optimal rendezvous trajectories of a controlled spacecraft and a tumbling object,”Journal of Guidance, Control, and dynamics, Vol. 34, No. 4, 2011, pp. 1239–1252

    work page 2011

  3. [3]

    On-orbit servicing: a new value proposition for satellite design and operation,

    Long, A. M., Richards, M. G., and Hastings, D. E., “On-orbit servicing: a new value proposition for satellite design and operation,”Journal of Spacecraft and Rockets, Vol. 44, No. 4, 2007, pp. 964–976

    work page 2007

  4. [4]

    A review of space robotics technologies for on-orbit servicing,

    Flores-Abad, A., Ma, O., Pham, K., and Ulrich, S., “A review of space robotics technologies for on-orbit servicing,”Progress in aerospace sciences, Vol. 68, 2014, pp. 1–26

    work page 2014

  5. [5]

    Relative computer vision-based navigation for small inspection spacecraft,

    Tweddle, B. E., and Saenz-Otero, A., “Relative computer vision-based navigation for small inspection spacecraft,”Journal of guidance, control, and dynamics, Vol. 38, No. 5, 2015, pp. 969–978. 20

    work page 2015

  6. [6]

    Information-based guidance and control architecture for multi-spacecraft on-orbit inspection,

    Nakka, Y. K., Hönig, W., Choi, C., Harvard, A., Rahmani, A., and Chung, S.-J., “Information-based guidance and control architecture for multi-spacecraft on-orbit inspection,”Journal of Guidance, Control, and Dynamics, Vol. 45, No. 7, 2022, pp. 1184–1201

    work page 2022

  7. [7]

    Peer-to-peer refueling strategy using low-thrust propulsion,

    Dutta, A., Arora, N., and Russell, R. P., “Peer-to-peer refueling strategy using low-thrust propulsion,”Journal of Spacecraft and Rockets, Vol. 49, No. 5, 2012, pp. 944–954

    work page 2012

  8. [8]

    Towards a standardized grasping and refuelling on-orbit servicing for geo spacecraft,

    Medina, A., Tomassini, A., Suatoni, M., Avilés, M., Solway, N., Coxhill, I., Paraskevas, I. S., Rekleitis, G., Papadopoulos, E., Krenn, R., et al., “Towards a standardized grasping and refuelling on-orbit servicing for geo spacecraft,”Acta Astronautica, Vol. 134, 2017, pp. 1–10

    work page 2017

  9. [9]

    On-orbit upgrade and repair: The hubble space telescope example,

    Hastings, D. E., and Joppin, C., “On-orbit upgrade and repair: The hubble space telescope example,”Journal of spacecraft and rockets, Vol. 43, No. 3, 2006, pp. 614–625

    work page 2006

  10. [10]

    The case for robotic on-orbit servicing of spacecraft: Spacecraft reliability is a myth,

    Ellery, A., Kreisel, J., and Sommer, B., “The case for robotic on-orbit servicing of spacecraft: Spacecraft reliability is a myth,” Acta Astronautica, Vol. 63, No. 5-6, 2008, pp. 632–648

    work page 2008

  11. [11]

    In-orbit assembly mission for the space solar power station,

    Cheng, Z., Hou, X., Zhang, X., Zhou, L., Guo, J., and Song, C., “In-orbit assembly mission for the space solar power station,” Acta Astronautica, Vol. 129, 2016, pp. 299–308

    work page 2016

  12. [12]

    Buildinglargetelescopes inorbitusing smallsatellites,

    Saunders, C., Lobb, D., Sweeting, M., andGao, Y., “Buildinglargetelescopes inorbitusing smallsatellites,”Acta Astronautica, Vol. 141, 2017, pp. 183–195

    work page 2017

  13. [13]

    Controlling the growth of future LEO debris populations with active debris removal,

    Liou, J.-C., Johnson, N. L., and Hill, N., “Controlling the growth of future LEO debris populations with active debris removal,” Acta Astronautica, Vol. 66, No. 5-6, 2010, pp. 648–653

    work page 2010

  14. [14]

    Active debris removal: Recent progress and current trends,

    Bonnal, C., Ruault, J.-M., and Desjean, M.-C., “Active debris removal: Recent progress and current trends,”Acta Astronautica, Vol. 85, 2013, pp. 51–60

    work page 2013

  15. [15]

    The kessler syndrome: implications to future space operations,

    Kessler, D. J., Johnson, N. L., Liou, J., and Matney, M., “The kessler syndrome: implications to future space operations,” Advances in the Astronautical Sciences, Vol. 137, No. 8, 2010, p. 2010

    work page 2010

  16. [16]

    On-orbitroboticgraspingofaspentrocketstage: Graspstabilityanalysisandexperimental results,

    Mavrakis,N.,Hao,Z.,andGao,Y.,“On-orbitroboticgraspingofaspentrocketstage: Graspstabilityanalysisandexperimental results,”Frontiers in Robotics and AI, Vol. 8, 2021, p. 652681

    work page 2021

  17. [17]

    Target Grasp-conditioned Whole-body Control of Satellite Manipulators using Neural Fields with MPC,

    Sah, R., Srivastava, R., Vatsal, V., Lima, R., and Das, K., “Target Grasp-conditioned Whole-body Control of Satellite Manipulators using Neural Fields with MPC,”2024 IEEE Aerospace Conference, IEEE, 2024, pp. 1–12

    work page 2024

  18. [18]

    Leveraging neural radiance fields for pose estimation of an unknown space object during proximity operations,

    Legrand, A., Detry, R., and De Vleeschouwer, C., “Leveraging neural radiance fields for pose estimation of an unknown space object during proximity operations,”IEEE International Conference on Space Robotics (ISpaRo) Workshop, available at arXiv:2405.12728, 2024

    work page arXiv 2024

  19. [19]

    Spacecraft state estimation using neural radiance fields,

    Heintz, A. M., and Peck, M., “Spacecraft state estimation using neural radiance fields,”Journal of Guidance, Control, and Dynamics, Vol. 46, No. 8, 2023, pp. 1596–1609. 21

    work page 2023

  20. [20]

    Nerf: Representing scenes as neural radiance fields for view synthesis,

    Mildenhall, B., Srinivasan, P. P., Tancik, M., Barron, J. T., Ramamoorthi, R., and Ng, R., “Nerf: Representing scenes as neural radiance fields for view synthesis,”Communications of the ACM, Vol. 65, No. 1, 2021, pp. 99–106

    work page 2021

  21. [21]

    3D Gaussian splatting for real-time radiance field rendering

    Kerbl, B., Kopanas, G., Leimkühler, T., and Drettakis, G., “3D Gaussian splatting for real-time radiance field rendering.”ACM Trans. Graph., Vol. 42, No. 4, 2023, pp. 139–1

    work page 2023

  22. [22]

    Vision-based neural scene representations for spacecraft,

    Mergy, A., Lecuyer, G., Derksen, D., and Izzo, D., “Vision-based neural scene representations for spacecraft,”Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2021, pp. 2002–2011

    work page 2021

  23. [23]

    Characterizingsatellitegeometryviaaccelerated3DGaussian splatting,

    Nguyen,V.M.,Sandidge,E.,Mahendrakar,T.,andWhite,R.T.,“Characterizingsatellitegeometryviaaccelerated3DGaussian splatting,”Aerospace, Vol. 11, No. 3, 2024, p. 183

    work page 2024

  24. [24]

    Sources of uncertainty in 3d scene reconstruction,

    Klasson, M., Mereu, R., Kannala, J., and Solin, A., “Sources of uncertainty in 3d scene reconstruction,”European Conference on Computer Vision, Springer, 2024, pp. 271–289

    work page 2024

  25. [25]

    Nerf in the wild: Neural radiance fields for unconstrained photo collections,

    Martin-Brualla, R., Radwan, N., Sajjadi, M. S., Barron, J. T., Dosovitskiy, A., and Duckworth, D., “Nerf in the wild: Neural radiance fields for unconstrained photo collections,”Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, 2021, pp. 7210–7219

    work page 2021

  26. [26]

    Graf: Generativeradiancefieldsfor3d-awareimagesynthesis,

    Schwarz,K.,Liao,Y.,Niemeyer,M.,andGeiger,A.,“Graf: Generativeradiancefieldsfor3d-awareimagesynthesis,”Advances in neural information processing systems, Vol. 33, 2020, pp. 20154–20166

    work page 2020

  27. [27]

    3d reconstruction of non-cooperative resident space objects using instant ngp-accelerated nerf and d-nerf,

    Caruso, B., Mahendrakar, T., Nguyen, V. M., White, R. T., and Steffen, T., “3d reconstruction of non-cooperative resident space objects using instant ngp-accelerated nerf and d-nerf,”arXiv preprint arXiv:2301.09060, 2023

    work page arXiv 2023

  28. [28]

    Neural Field-Based Space Target 3D Reconstruction with Predicted Depth Priors,

    Fu, T., Zhou, Y., Wang, Y., Liu, J., Zhang, Y., Kong, Q., and Chen, B., “Neural Field-Based Space Target 3D Reconstruction with Predicted Depth Priors,”Aerospace, Vol. 11, No. 12, 2024, p. 997

    work page 2024

  29. [29]

    Joint attitude estimation and 3D neural reconstruction of non-cooperative space objects,

    Forray, C., Delporte, P., Delaygue, N., Genin, F., and Derksen, D., “Joint attitude estimation and 3D neural reconstruction of non-cooperative space objects,”arXiv preprint arXiv:2506.20638, 2025

    work page arXiv 2025

  30. [30]

    Spacecraft-NeRF: High-Fidelity Reconstruction of Spacecraft by Neural Radiance Field Based Implicit Representation,

    Liu, Y., Sun, Z., Zhang, L., Xi, L., Dong, W., and Deng, F., “Spacecraft-NeRF: High-Fidelity Reconstruction of Spacecraft by Neural Radiance Field Based Implicit Representation,”IEEE Transactions on Aerospace and Electronic Systems, 2025, pp. 1–13. https://doi.org/10.1109/TAES.2025.3539273

    work page doi:10.1109/taes.2025.3539273 2025

  31. [31]

    Instant neural graphics primitives with a multiresolution hash encoding,

    Müller, T., Evans, A., Schied, C., and Keller, A., “Instant neural graphics primitives with a multiresolution hash encoding,” ACM transactions on graphics (TOG), Vol. 41, No. 4, 2022, pp. 1–15

    work page 2022

  32. [32]

    D-nerf: Neural radiance fields for dynamic scenes,

    Pumarola, A., Corona, E., Pons-Moll, G., and Moreno-Noguer, F., “D-nerf: Neural radiance fields for dynamic scenes,” Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, 2021, pp. 10318–10327

    work page 2021

  33. [33]

    Pose estimation and neural implicit reconstruction towards non-cooperative spacecraft without offline prior information,

    Han, B., Wang, C., Zhang, X., Zhao, Z., Zhai, Z., Liu, J., Liu, N., and Chen, X., “Pose estimation and neural implicit reconstruction towards non-cooperative spacecraft without offline prior information,”IEEE Transactions on Aerospace and Electronic Systems, 2024. 22

    work page 2024

  34. [34]

    A neural radiance fields method for 3D reconstruction of space target,

    Xu, Q., Hu, M., Fang, Y., and Zhang, X., “A neural radiance fields method for 3D reconstruction of space target,”Advances in Space Research, Vol. 75, No. 9, 2025, pp. 6924–6943

    work page 2025

  35. [35]

    On Optimal Observation Orbits for Learning Gaussian Splatting-Based 3D Models of Unknown Resident Space Objects,

    Issitt, A., Mahendrakar, T., Alvarez, A., White, R. T., and Sizemore, A., “On Optimal Observation Orbits for Learning Gaussian Splatting-Based 3D Models of Unknown Resident Space Objects,”AIAA SCITECH 2025 Forum, 2025, p. 1780

    work page 2025

  36. [36]

    Improved 3D Gaussian Splatting of Unknown Spacecraft Structure Using Space Environment Illumination Knowledge,

    Park, T. H., and D’Amico, S., “Improved 3D Gaussian Splatting of Unknown Spacecraft Structure Using Space Environment Illumination Knowledge,”2025 International Conference on Space Robotics (iSpaRo), IEEE, 2025

    work page 2025

  37. [37]

    OntheGeometricAccuracyofImplicitandPrimitive-basedRepresentations Derived from View Rendering Constraints,

    DeSmijter,E.,Detry,R.,andDeVleeschouwer,C.,“OntheGeometricAccuracyofImplicitandPrimitive-basedRepresentations Derived from View Rendering Constraints,”arXiv preprint arXiv:2509.10241, 2025

    work page arXiv 2025

  38. [38]

    3D convex splatting: Radiance field rendering with 3D smooth convexes,

    Held, J., Vandeghen, R., Hamdi, A., Deliege, A., Cioppa, A., Giancola, S., Vedaldi, A., Ghanem, B., and Van Droogenbroeck, M., “3D convex splatting: Radiance field rendering with 3D smooth convexes,”Proceedings of the Computer Vision and Pattern Recognition Conference, 2025, pp. 21360–21369

    work page 2025

  39. [39]

    DreamSat: Towards a General 3D Model for Novel View Synthesis of Space Objects,

    Mathihalli, N., Wei, A., Lavezzi, G., Siew, P. M., Rodriguez-Fernandez, V., Urrutxua, H., and Linares, R., “DreamSat: Towards a General 3D Model for Novel View Synthesis of Space Objects,”arXiv preprint arXiv:2410.05097, 2024

    work page arXiv 2024

  40. [40]

    Rapid abstraction of spacecraft 3d structure from single 2d image,

    Park, T. H., and D’Amico, S., “Rapid abstraction of spacecraft 3d structure from single 2d image,”AIAA SCITECH 2024 Forum, 2024, p. 2768

    work page 2024

  41. [41]

    Improving Zero-Shot Abstraction of Unknown Spacecraft 3D Shape as Primitive Assembly,

    Park, T. H., Bates, E., and D’Amico, S., “Improving Zero-Shot Abstraction of Unknown Spacecraft 3D Shape as Primitive Assembly,”International Workshop on Satellite Constellations and Formation Flying, 2024

    work page 2024

  42. [42]

    Removing ambiguities in concurrent monocular single-shot spacecraft shape and pose estimation using a deep neural network,

    Bates, E., and D’Amico, S., “Removing ambiguities in concurrent monocular single-shot spacecraft shape and pose estimation using a deep neural network,”2025 AAS GNC conference (Feb 2025), 2025

    work page 2025

  43. [43]

    Fast Learning of Non-Cooperative SpaceCraft 3D Models Through Primitive Initialization,

    Huc, P. F., Bates, E., and D’Amico, S., “Fast Learning of Non-Cooperative SpaceCraft 3D Models Through Primitive Initialization,”arXiv preprint arXiv:2507.19459, 2025

    work page arXiv 2025

  44. [44]

    Domain generalization for 6D pose estimation through NeRF-based image synthesis,

    Legrand, A., Detry, R., and De Vleeschouwer, C., “Domain generalization for 6D pose estimation through NeRF-based image synthesis,”arXiv preprint arXiv:2407.10762, 2024

    work page arXiv 2024

  45. [45]

    NeRF-based Visualization of 3D Cues Supporting Data-Driven Spacecraft Pose Estimation,

    Legrand, A., Detry, R., and De Vleeschouwer, C., “NeRF-based Visualization of 3D Cues Supporting Data-Driven Spacecraft Pose Estimation,”2025 International Conference on Space Robotics (iSpaRo), IEEE, 2025

    work page 2025

  46. [46]

    imap: Implicit mapping and positioning in real-time,

    Sucar, E., Liu, S., Ortiz, J., and Davison, A. J., “imap: Implicit mapping and positioning in real-time,”Proceedings of the IEEE/CVF international conference on computer vision, 2021, pp. 6229–6238

    work page 2021

  47. [47]

    Nice-slam: Neural implicit scalable encoding for slam,

    Zhu, Z., Peng, S., Larsson, V., Xu, W., Bao, H., Cui, Z., Oswald, M. R., and Pollefeys, M., “Nice-slam: Neural implicit scalable encoding for slam,”Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, 2022, pp. 12786–12796. 23

    work page 2022

  48. [48]

    Nerf-slam: Real-time dense monocular slam with neural radiance fields,

    Rosinol, A., Leonard, J. J., and Carlone, L., “Nerf-slam: Real-time dense monocular slam with neural radiance fields,”2023 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), IEEE, 2023, pp. 3437–3444

    work page 2023

  49. [49]

    Object-centric Reconstruction and Tracking of Dynamic Unknown Objects Using 3D Gaussian Splatting,

    Barad, K. R., Richard, A., Dentler, J., Olivares-Mendez, M., and Martinez, C., “Object-centric Reconstruction and Tracking of Dynamic Unknown Objects Using 3D Gaussian Splatting,”2024 International Conference on Space Robotics (iSpaRo), IEEE, 2024, pp. 202–209

    work page 2024

  50. [50]

    Nope-nerf: Optimising neural radiance field with no pose prior,

    Bian, W., Wang, Z., Li, K., Bian, J.-W., and Prisacariu, V. A., “Nope-nerf: Optimising neural radiance field with no pose prior,” Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2023, pp. 4160–4169

    work page 2023

  51. [51]

    CBARF: cascaded bundle-adjusting neural radiance fields from imperfect camera poses,

    Fu, H., Yu, X., Li, L., and Zhang, L., “CBARF: cascaded bundle-adjusting neural radiance fields from imperfect camera poses,” IEEE Transactions on Multimedia, Vol. 26, 2024, pp. 9304–9315

    work page 2024

  52. [52]

    inerf: Inverting neural radiance fields for pose estimation,

    Yen-Chen, L., Florence, P., Barron, J. T., Rodriguez, A., Isola, P., and Lin, T.-Y., “inerf: Inverting neural radiance fields for pose estimation,”2021 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), IEEE, 2021, pp. 1323–1330

    work page 2021

  53. [53]

    Parallel inversion of neural radiance fields for robust pose estimation,

    Lin, Y., Müller, T., Tremblay, J., Wen, B., Tyree, S., Evans, A., Vela, P. A., and Birchfield, S., “Parallel inversion of neural radiance fields for robust pose estimation,”arXiv preprint arXiv:2210.10108, 2022

    work page arXiv 2022

  54. [54]

    Barf: Bundle-adjusting neural radiance fields,

    Lin, C.-H., Ma, W.-C., Torralba, A., and Lucey, S., “Barf: Bundle-adjusting neural radiance fields,”Proceedings of the IEEE/CVF international conference on computer vision, 2021, pp. 5741–5751

    work page 2021

  55. [55]

    K-planes: Explicit radiance fields in space, time, and appearance,

    Fridovich-Keil, S., Meanti, G., Warburg, F. R., Recht, B., and Kanazawa, A., “K-planes: Explicit radiance fields in space, time, and appearance,”Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2023, pp. 12479–12488

    work page 2023

  56. [56]

    Adaptiveneural-network-basedunscentedkalmanfilterforrobustposetrackingofnoncooperative spacecraft,

    Park,T.H.,andD’Amico,S.,“Adaptiveneural-network-basedunscentedkalmanfilterforrobustposetrackingofnoncooperative spacecraft,”Journal of Guidance, Control, and Dynamics, Vol. 46, No. 9, 2023, pp. 1671–1688

    work page 2023

  57. [57]

    SPEED+: Next-generation dataset for spacecraft pose estimation across domain gap,

    Park, T. H., Märtens, M., Lecuyer, G., Izzo, D., and D’Amico, S., “SPEED+: Next-generation dataset for spacecraft pose estimation across domain gap,”2022 IEEE aerospace conference (AERO), IEEE, 2022, pp. 1–15

    work page 2022

  58. [58]

    Autonomous formation flying for the PRISMA mission,

    Gill, E., D’Amico, S., and Montenbruck, O., “Autonomous formation flying for the PRISMA mission,”Journal of Spacecraft and Rockets, Vol. 44, No. 3, 2007, pp. 671–681. https://doi.org/10.2514/1.23015

    work page doi:10.2514/1.23015 2007

  59. [59]

    Robotic testbed for rendezvous and optical navigation: Multi-source calibration and machine learning use cases,

    Park, T. H., Bosse, J., and D’Amico, S., “Robotic testbed for rendezvous and optical navigation: Multi-source calibration and machine learning use cases,”arXiv preprint arXiv:2108.05529, 2021

    work page arXiv 2021

  60. [60]

    Image quality assessment: from error visibility to structural similarity,

    Wang, Z., Bovik, A. C., Sheikh, H. R., and Simoncelli, E. P., “Image quality assessment: from error visibility to structural similarity,”IEEE transactions on image processing, Vol. 13, No. 4, 2004, pp. 600–612

    work page 2004

  61. [61]

    Ano-referenceobjectiveimagesharpnessmetricbasedonthenotionofjustnoticeableblur(JNB),

    Ferzli,R.,andKaram,L.J.,“Ano-referenceobjectiveimagesharpnessmetricbasedonthenotionofjustnoticeableblur(JNB),” IEEE transactions on image processing, Vol. 18, No. 4, 2009, pp. 717–728. 24

    work page 2009