Pose estimation for non-cooperative spacecraft rendezvous using convolutional neural networks
Open accessAbstract
On-board estimation of the pose of an uncooperative target spacecraft is an essential task for future on-orbit servicing and close-proximity formation flying missions. However, two issues hinder reliable on-board monocular vision based pose estimation: robustness to illumination conditions due to a lack of reliable visual features and scarcity of image datasets required for training and benchmarking. To address these two issues, this work details the design and validation of a monocular vision based pose determination architecture for spaceborne applications. The primary contribution to the state-of-the-art of this work is the introduction of a novel pose determination method based on Convolutional Neural Networks (CNN) to provide an initial guess of the pose in real-time on-board. The method involves discretizing the pose space and training the CNN with images corresponding to the resulting pose labels. Since reliable training of the CNN requires massive image datasets and computational resources, the parameters of the CNN must be determined prior to the mission with synthetic imagery. Moreover, reliable training of the CNN requires datasets that appropriately account for noise, color, and illumination characteristics expected in orbit. Therefore, the secondary contribution of this work is the introduction of an image synthesis pipeline, which is tailored to generate high fidelity images of any spacecraft 3D model. In contrast to prior techniques demonstrated for close-range pose determination of spacecraft, the proposed architecture relies on neither hand-engineered image features nor a-priori relative state information. Hence, the proposed technique is scalable to spacecraft of different structural and physical properties as well as robust to the dynamic illumination conditions of space. Through metrics measuring classification and pose accuracy, it is shown that the presented architecture has desirable robustness and scalable properties. Therefore, the proposed technique can be used to augment the current state-of-the-art monocular vision-based pose estimation techniques used in spaceborne applications.
The claims and statistics of this paper haven’t been checked yet.
Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.
Integrity signals
Registry
Cites
Statistics
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
Pose errors of the CNN in [17] considerably increase when the network is tested with real images.
“This can be clearly seen in Table 6, in which the pose errors considerably increase when the network is tested with real images.”
From Review of the robustness and applicability of monocular pose estimation systems for relative navigation with an uncooperative spacecraft (2019) · cites this paper as [17]
The Experiments section says the CNN was evaluated on imagery imitating actual PRISMA space imagery, not on real images. The Conclusions section explicitly says testing the networks on actual space imagery remains to be done, so the paper cannot report an increase in pose errors on real-image tests. No page number is available in our copy. (LLM, conf=0.89)
Checked by a reasoning model — may be wrong.
Experimental — accuracy not yet measuredThe scheme proposed in [56] returns better pose estimates than the AlexNet scheme while decreasing the size of the training set, and has comparable accuracy in 2D bounding box detection.
“The scheme proposed in [56] proved to return better pose estimates than the AlexNet scheme while at the same decreasing the size of the training set, as well as a comparable accuracy in the 2D bounding box detection compared to the architecture in [56].”
From Review of the robustness and applicability of monocular pose estimation systems for relative navigation with an uncooperative spacecraft (2019) · cites this paper as [56]
The Convolutional Neural Network section says the proposed networks adopt AlexNet’s architecture, rather than outperforming a separate AlexNet scheme. The Conclusions state that training-set size correlates with accuracy and call for larger datasets, contrary to the claimed improvement from a smaller training set. Our copy does not report a comparison of 2D bounding-box detection accuracy. (LLM, conf=0.86)
Checked by a reasoning model — may be wrong.
Experimental — accuracy not yet measuredThe TANGO satellite model has a coarse dimension of 80 × 75 × 32 cm and leads to an operation range of approximately 10 m.
“The TANGO satellite [145] model used in the multiple datasets [19,20,51,51,85,132] has a coarse dimension of 80 × 75 × 32 cm will lead to the operation range of ∼10 m.”
From A survey on deep learning-based monocular spacecraft pose estimation: Current state, limitations and prospects (2023) · cites this paper as [85]
The Synthetic Dataset Creation section (page unavailable) says the paper generated images of the Tango spacecraft, and the dataset description lists camera-to-target distances of 8–13 m, including 10 m. Our copy does not give dimensions of 80 × 75 × 32 cm or say that the satellite’s size leads to an approximately 10 m operating range. (LLM, conf=0.85)
Experimental — accuracy not yet measuredSharma et al. later proposed doing position estimation based on bounding box detection and orientation estimation based on soft classification.
“but later they [29] proposed doing position estimation based on bounding box detection and orientation estimation based on soft classification.”
From Deep Learning for Spacecraft Pose Estimation from Photorealistic Rendering (2020) · cites this paper as [29]
The Introduction (page unavailable) says the architecture requires a sliding-window approach to locate the spacecraft in the image and uses bearing information to reconstruct its three-dimensional position; it does not describe position estimation based on bounding-box detection. The Type 2 section (page unavailable) describes softmax probabilities for pose labels, supporting classification-based pose estimation, but not a separate soft-classification method specifically for orientation. (LLM, conf=0.86)
Experimental — accuracy not yet measuredIn [16], initial conditions for relative state were assumed from separate results of the pose initialization subsystem, without modeling the interface between initial pose estimation and the filter, and no SIL/HIL tests were conducted.
“Moreover, the initial conditions for the relative state in [16] were assumed from the separate results of the pose initialization subsystem, without modeling the interface between the initial pose estimation and the filter itself, and no SIL/HIL tests were conducted.”
From Review of the robustness and applicability of monocular pose estimation systems for relative navigation with an uncooperative spacecraft (2019) · cites this paper as [16]
The Conclusions (page unavailable) describe the CNN as a potential initializer for other pose-determination algorithms, not as an initializer connected to a filter. The Type 1 and Type 2 experimental sections (pages unavailable) report image-dataset tests, consistent with the claim that no SIL/HIL tests were conducted. Our copy does not support the more specific assertion that relative-state initial conditions were assumed from separate pose-initialization results. (LLM, conf=0.78)
Experimental — accuracy not yet measured
Internal dynamics of the filter improve accuracy of the predicted relative state from measurements and allow more robust pose tracking.
“Furthermore, the internal dynamics of the filter improve the accuracy of the predicted relative state from measurements and allow a more robust pose tracking.”
From Review of the robustness and applicability of monocular pose estimation systems for relative navigation with an uncooperative spacecraft (2019) · cites this paper as [16]
The Introduction (page unavailable) says conventional pose-determination methods use filtering techniques for closed-loop pose tracking. It does not say that a filter’s internal dynamics improve predicted relative-state accuracy or make tracking more robust; the paper’s reported accuracy results concern its CNN approach. (LLM, conf=0.78)
Experimental — accuracy not yet measuredSince training in [17] and [56] used relative distances from 3 to 50 meters as labels, the estimation system for close-proximity operations down to docking could not be validated.
“Notice also that, since the training in [17] and [56] has been performed with relative distances from 3 up to 50 meters as labels, the estimation system for close-proximity 570 operations down to docking could not be validated.”
From Review of the robustness and applicability of monocular pose estimation systems for relative navigation with an uncooperative spacecraft (2019) · cites this paper as [56]
Table 1 (section labeled 4. CONCLUSIONS; page unavailable) lists pose-label distances starting at 3 m, with no labels for distances closer to docking. But its largest listed label distance is 12 m, not 50 m; the nearby value 50 is a separate image-generation parameter, not a distance in meters. The paper does not explicitly state that docking-range validation was impossible. (LLM, conf=0.87)
Experimental — accuracy not yet measuredSince training in [17] and [56] used relative distances from 3 to 50 meters as labels, the estimation system for close-proximity operations down to docking could not be validated.
“Notice also that, since the training in [17] and [56] has been performed with relative distances from 3 up to 50 meters as labels, the estimation system for close-proximity 570 operations down to docking could not be validated.”
From Review of the robustness and applicability of monocular pose estimation systems for relative navigation with an uncooperative spacecraft (2019) · cites this paper as [17]
Table 1 (section labeled 4. CONCLUSIONS; page not provided) lists pose-label distances starting at 3 m, with none below 3 m, which supports the limitation concerning validation down to docking. But this paper’s listed labels extend only to 12 m, not 50 m; the claim also cites another source, which is not evidence for this verdict. (LLM, conf=0.87)
Experimental — accuracy not yet measuredThe approach to position fails when part of object is outside the field of view.
“Although, the approach to position fails when part of object is outside the field of view, the orientation estimation has its merits.”
From Deep Learning for Spacecraft Pose Estimation from Photorealistic Rendering (2020) · cites this paper as [29]
The Synthetic Dataset Creation passage describes images in which some spacecraft features may fall outside the camera’s viewing cone. The Introduction says the architecture estimates range and needs a sliding-window detection step to reconstruct three-dimensional position. Neither passage establishes that position estimation fails when part of the spacecraft is outside the field of view. (LLM, conf=0.76)
Experimental — accuracy not yet measuredThe ESA challenge dataset is just made of montages of real images with basic OpenGL renderings of a satellite.
“The same work [29] introduced the dataset used in the ESA challenge, which is just made of montages of real images with basic OpenGL renderings of a satellite.”
From Deep Learning for Spacecraft Pose Estimation from Photorealistic Rendering (2020) · cites this paper as [29]
The Synthetic Dataset Creation section (page unavailable) says the authors created satellite images using OpenGL, supporting that part of the claim. It describes them as synthetic images, not a dataset made of montages of real images; the montages mentioned in the paper are figures displaying examples. Our copy does not identify the dataset as one used in an ESA challenge. (LLM, conf=0.86)
Experimental — accuracy not yet measuredSharma et al. were the first to propose using CNNs for spacecraft pose estimation based on hard viewpoint classification.
“Sharma et al. [28] were the first to propose using CNNs for spacecraft pose estimation based on hard viewpoint classification”
From Deep Learning for Spacecraft Pose Estimation from Photorealistic Rendering (2020) · cites this paper as [28]
The CNN-based algorithm in [17] has been extensively tested against the number of synthetic images used in training, different levels of image noise, and the amount of displacement of the target from the center of the image plane.
“The CNN-based algorithm in [17] has been extensively tested against the number of synthetic images used in the training, different levels of image noise and the amount of displacement of the target from the center of the image plane, which has 550 not been tested in the validation of other pose estimation algorithms.”
From Review of the robustness and applicability of monocular pose estimation systems for relative navigation with an uncooperative spacecraft (2019) · cites this paper as [17]
A feature-based algorithm combined with CNN-based pose estimation, which provides a coarse initial guess, could increase robustness of pose initialization in scenarios where IP fails to extract target features from the image background.
“As mentioned in [17], a feature-based algorithm with a CNN-based pose estimation, which provides a coarse initial guess, could increase the robustness of the pose initializa-575 tion with respect to scenarios in which the IP fails in extracting the target features from the image background.”
From Review of the robustness and applicability of monocular pose estimation systems for relative navigation with an uncooperative spacecraft (2019) · cites this paper as [17]
The ESA & Stanford benchmark dataset (SPEED) contains both synthetic and real images of a mockup model of one satellite used in a flight mission, named PRISMA.
“The latter contains both synthetic and real images with 1920 × 1200 px, generated in [29], of a mockup model of one satellite used in a flight mission, named PRISMA [44].”
From Deep Learning for Spacecraft Pose Estimation from Photorealistic Rendering (2020) · cites this paper as [29]
SPEED comprises 15,000 synthetic images and 300 simulated images captured from TRON.
“Specifically, the dataset comprises 15,000 synthetic images rendered with OpenGLbased Optical Stimulator (OS) camera emulator software [19, 20] of the Stanford's Space Rendezvous Laboratory (SLAB) multi-Satellite Software Simulator ( S 3 ) [21] and 300 simulated images captured from the robotic Testbed for Rendezvous and Optical Navigation (TRON) at SLAB.”
From SPEED+: Next-Generation Dataset for Spacecraft Pose Estimation across Domain Gap (2022) · cites this paper as [19]
The Synthetic Dataset Creation section describes this paper’s own OpenGL-rendered image sets, including a superset of 500,000 images; it does not describe SPEED. Searches for SPEED, the 15,000/300-image counts, TRON, and the S3 simulator found no passage establishing the cited claim. The different counts for this paper’s own images do not contradict the composition of SPEED. (LLM, conf=0.76)
Experimental — accuracy not yet measuredIn both implementations [12] and [16], filter dynamics were highly simplified and no perturbation models were included.
“However, in both implementations the filter dynamics were highly simplified and no perturbation models were included.”
From Review of the robustness and applicability of monocular pose estimation systems for relative navigation with an uncooperative spacecraft (2019) · cites this paper as [16]
The Synthetic Dataset Creation section describes generating images with varying noise levels, and the Type 1 section evaluates CNN classification; neither addresses filter dynamics or perturbation models. Searches for “filter dynamics,” “Kalman filter,” and “perturbation” found no passage in our copy establishing the attributed claim. (LLM, conf=0.73)
Experimental — accuracy not yet measuredThe scheme proposed in [56] is robust with respect to Earth in the background.
“Furthermore, it proved to be robust with respect to the Earth in 565 the background.”
From Review of the robustness and applicability of monocular pose estimation systems for relative navigation with an uncooperative spacecraft (2019) · cites this paper as [56]
The Synthetic Dataset Creation section describes tests with image noise and off-center spacecraft, while the Conclusions discuss robustness to noise and dynamic illumination. Neither passage addresses an Earth background, and searches for “Earth,” “Earth in the background,” and background clutter found no such test or finding in our copy. Page numbers were not available. (LLM, conf=0.72)
Experimental — accuracy not yet measuredThe performance of the scheme in [56] drops off at relatively close distances where the target is not fully in the camera FoV and during poor illumination conditions close to eclipse, due to inaccurate box detections.
“However, its performance was found to drop-off at relatively close distances for which the target is not fully in the camera FoV as well as during poor illumination conditions close to eclipse, due to inaccurate box detections.”
From Review of the robustness and applicability of monocular pose estimation systems for relative navigation with an uncooperative spacecraft (2019) · cites this paper as [56]
Synthetic Dataset Creation describes images with target features outside the camera’s viewing cone, and the Introduction discusses illumination challenges for conventional methods. Neither section reports that this scheme’s performance drops at close range or near eclipse because of inaccurate box detections; searches for bounding boxes, eclipse, and field of view found no such result. (LLM, conf=0.72)
Experimental — accuracy not yet measured
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
Cites
Very Deep Convolutional Networks for Large-Scale Image Recognition
Simonyan, K., & Zisserman, A. (2014) · arXiv · doi:10.48550/arxiv.1409.1556 · Cited by 74,248 · see citing works
ImageNet Large Scale Visual Recognition Challenge
Russakovsky, O., Deng, J., Su, H., Krause, J., Satheesh, S., Ma, S., Huang, Z., Karpathy, A., Khosla, A., Bernstein, M., et al. (2015) · International Journal of Computer Vision · doi:10.1007/s11263-015-0816-y · Cited by 39,380 · see citing works
Fischler, M. A., & Bolles, R. C. (1981) · Communications of the ACM · doi:10.1145/358669.358692 · Cited by 25,893 · see citing works
Inception-v4, Inception-ResNet and the Impact of Residual Connections on Learning
Szegedy, C., Ioffe, S., Vanhoucke, V., & Alemi, A. (2017) · Proceedings of the AAAI Conference on Artificial Intelligence · doi:10.1609/aaai.v31i1.11231 · Cited by 4,636 · see citing works
EPnP: An Accurate O(n) Solution to the PnP Problem
Lepetit, V., Moreno-Noguer, F., & Fua, P. (2009) · International Journal of Computer Vision · doi:10.1007/s11263-008-0152-6 · Cited by 3,298 · see citing works
DeepPose: Human Pose Estimation via Deep Neural Networks
Toshev, A., & Szegedy, C. (2014) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr.2014.214 · Cited by 3,288 · see citing works
OverFeat: Integrated Recognition, Localization and Detection using Convolutional Networks
Sermanet, P., Eigen, D., Zhang, X., Mathieu, M., Fergus, R., & LeCun, Y. (2013) · arXiv · doi:10.48550/arxiv.1312.6229 · Cited by 2,680 · see citing works
PoseNet: A Convolutional Network for Real-Time 6-DOF Camera Relocalization
Kendall, A., Grimes, M., & Cipolla, R. (2015) · 2015 IEEE International Conference on Computer Vision (ICCV) · doi:10.1109/iccv.2015.336 · Cited by 2,489 · see citing works
Joint Training of a Convolutional Network and a Graphical Model for Human Pose EstimationPreprint
Tompson, J., Jain, A., LeCun, Y., & Bregler, C. (2014) · arXiv (Cornell University) · doi:10.48550/arxiv.1406.2984 · Cited by 1,317 · see citing works
An Analysis of Deep Neural Network Models for Practical Applications
Canziani, A., Paszke, A., & Culurciello, E. (2016) · arXiv · doi:10.48550/arxiv.1605.07678 · Cited by 969 · see citing works
Spaceborne Autonomous Formation-Flying Experiment on the PRISMA Mission
D'Amico, S., Ardaens, J.-S., & Larsson, R. (2012) · Journal of Guidance, Control, and Dynamics · doi:10.2514/1.55638 · Cited by 173 · see citing works
Self-citationThe Restore-L Servicing Mission
Reed, B. B., Smith, R. C., Naasz, B. J., Pellegrino, J. F., & Bacon, C. E. (2016) · AIAA SPACE 2016 · doi:10.2514/6.2016-5478 · Cited by 119 · see citing works
Fast and Near-Optimal Guidance for Docking to Uncontrolled Spacecraft
Ventura, J., Ciarcià, M., Romano, M., et al. (2016) · Journal of Guidance Control and Dynamics · doi:10.2514/1.g001843 · Cited by 80 · see citing works
Real-time pose estimation of 3D objects from camera images using neural networks
Wunsch, P., WINKLER, S. S., & Hirzinger, G. (2002) · Proceedings - IEEE International Conference on Robotics and Automation/Proceedings · doi:10.1109/robot.1997.606781 · Cited by 42 · see citing works
Khotanzad, A. R., & Liou, J. J.-H. (1996) · IEEE Transactions on Neural Networks · doi:10.1109/72.508933 · Cited by 41 · see citing works
Castellani, L. T., Llorente, S., Ibarz, J. M. F., Ruiz, M., Mestreau-Garreau, A., Cropp, A., & Santovincenzo, A. (2013) · International Journal of Space Science and Engineering · doi:10.1504/ijspacese.2013.059268 · Cited by 33 · see citing works
Final commissioning of the PRISMA GPS navigation system
Ardaens, J.-S., D’Amico, S., & Montenbruck, O. (2012) · Journal of Aerospace Engineering Sciences and Applications · doi:10.7446/jaesa.0403.10 · Cited by 28 · see citing works
Self-citationAppearance learning for 3D pose detection of a satellite at close-range
Oumer, N. W., Kriegel, S., Ali, H., et al. (2017) · ISPRS Journal of Photogrammetry and Remote Sensing · doi:10.1016/j.isprsjprs.2017.01.002 · Cited by 10 · see citing works
Bowick, M. J., Cacciuto, A., Nelson, D. R., et al. (2002) · Physical Review Letters · doi:10.1103/physrevlett.89.249902 · Cited by 4 · see citing works
Flagged first, then by citations
Research works printed but not linked: 16
- [8]8. J. Ventura, "Autonomous Proximity Operations for Noncooperative Space Target," Ph.D. dissertation, Technical University of Munich, 2016.
- [9]9. A. Cropp and P. Palmer, "Pose Estimation and Relative Orbit Determination of a Nearby Target Microsatellite using Passive Imagery," 5th Cranfield Conference on Dynamics and Control of Systems and Structures in Space 2002, pp. 389-395, 2002.
- [10]10. M. Benn, "Vision Based Navigation Sensors for Space- craft Rendezvous and Docking," Ph.D. dissertation, Danish Technical University, 2010.
- [11]11. K. Kanani, A. Petit, E. Marchand, T. Chabot, and B. Gerber, "Vision Based Navigation for Debris Removal Missions," 63rd International Astronautical Congress, pp. 1-8, 2012. [Online]. Available: http: //hal.inria.fr/hal-00750610
- [12]12. S. Sharma and S. D'Amico, "Comparative Assessment of Techniques for Initial Pose Estimation using Monoc- ular Vision," Acta Astronautica, pp. 1-17, 2015.
- [14]14. S. D'Amico, M. Benn, and J. Jorgensen, "Pose Estima- tion of an Uncooperative Spacecraft," Proceedings of 5th International Conference on Spacecraft Formation Flying Missions and Technologies, no. 1, pp. 1-17, 2013.
- [15]15. S. Sharma and J. Ventura, "Robust Model-Based Monocular Pose Estimation for Noncooperative Space- craft Rendezvous," Journal of Spacecraft and Rockets, pp. 1-35, 2017.
- [19]19. A. R. Romano, "3D Pose Estimation Using Convo- lutional Neural Networks," Ph.D. dissertation, Escola Tècnica Superior d'Enginyeria Industrial de Barcelona, 2015.
- [20]20. H. Su, C. R. Qi, Y. Li, and L. J. Guibas, "Render for CNN: Viewpoint estimation in images using CNNs trained with rendered 3D model views," Proceedings of the IEEE International Conference on Computer Vision, vol. 11-18-Dece, pp. 2686-2694, 2016.
- [21]21. M. W. Wright and F. Fallside, "Object Pose Estimation by Neural Network," in International Conference on Image Processing and its Applications. Maastricht, Netherlands: IET, 1992, pp. 602-603.
- [27]27. A. Krizhevsky, I. Sutskever, and G. E. Hinton, "Ima- geNet Classification with Deep Convolutional Neural Networks," Advances In Neural Information Processing Systems, pp. 1-9, 2012.
- [28]28. B. Zhou, A. Lapedriza, J. Xiao, A. Torralba, and A. Oliva, "Learning Deep Features for Scene Recognition using Places Database," Advances in Neural Information Processing Systems 27, pp. 487-495, 2014. [Online]. Available: http://papers.nips.cc/paper/ 5349-learning-deep-features-for-scene-recognition-using-places-database. pdf
- [29]29. M. Bowick, A. Cacciuto, D. R. Nelson, and A. Travesset, "Crystalline Order on a Sphere and the Generalized Thomson Problem," Physical Review Letters, vol. 89, no. 24, p. 249902, 2002. [Online]. Available: https://link.aps.org/doi/10.1103/ PhysRevLett.89.249902
- [30]30. P. Bodin, R. Noteborn, R. Larsson, T. Karlsson, S. D'Amico, J.-S. Ardaens, M. Delpech, and J. C. Berges, "The prisma formation flying demonstrator: Overview and conclusions from the nominal mission," Advances in the Astronautical Sciences, vol. 144, pp. 441-460, 2012.
- [32]32. S. D'Amico, P. Bodin, M. Delpech, and R. Noteborn, "PRISMA," in Distributed Space Missions for Earth System Monitoring, M. D'Errico, Ed., 2013, pp. 599- 637.
- [36]36. G. E. Hinton, N. Srivastava, A. Krizhevsky, I. Sutskever, and R. R. Salakhutdinov, "Improving neural networks by preventing co-adaptation of feature detectors," pp. 1-18, 2012. [Online]. Available: http://arxiv.org/abs/ 1207.0580
Cited by
Sorted by citations
Spacecraft pose estimation dataset (speed)
Sharma, S., Park, T. H., & D'Amico, S. (2026) · Stanford Digital Repository · doi:10.25740/dz692fn7184 · Cited by 304
Self-citationSatellite Pose Estimation Challenge: Dataset, Competition Design, and Results
Kisantal, M., Sharma, S., Park, T. H., Izzo, D., Martens, M., & D'Amico, S. (2020) · IEEE Transactions on Aerospace and Electronic Systems · doi:10.1109/taes.2020.2989063 · Cited by 304 · see citing works
“The synthetic images of the Tango spacecraft are created using the camera emulator software of the Optical Stimulator [42], [43].”
Self-citationDeep Learning for Spacecraft Pose Estimation from Photorealistic Rendering
Proenca, P. F., & Gao, Y. (2020) · 2020 IEEE International Conference on Robotics and Automation (ICRA) · doi:10.1109/icra40945.2020.9197244 · Cited by 229 · see citing works
“Sharma et al. [28] were the first to propose using CNNs for spacecraft pose estimation based on hard viewpoint classification, but later they [29] proposed doing position estimation based on bounding box detection and orientation estimation based on soft classification.”
MethodNeural Network-Based Pose Estimation for Noncooperative Spacecraft Rendezvous
Sharma, S., & D'Amico, S. (2020) · IEEE Transactions on Aerospace and Electronic Systems · doi:10.1109/taes.2020.2999148 · Cited by 196 · see citing works
Self-citationPasqualetto Cassinis, L., Fonod, R., & Gill, E. (2019) · Progress in Aerospace Sciences · doi:10.1016/j.paerosci.2019.05.008 · Cited by 156 · see citing works
SPEED+: Next-Generation Dataset for Spacecraft Pose Estimation across Domain Gap
Park, T. H., Martens, M., Lecuyer, G., Izzo, D., & D'Amico, S. (2022) · 2022 IEEE Aerospace Conference (AERO) · doi:10.1109/aero53065.2022.9843439 · Cited by 170 · see citing works
Self-citationRobust multi-task learning and online refinement for spacecraft pose estimation across domain gap
Park, T. H., & D’Amico, S. (2024) · Advances in Space Research · doi:10.1016/j.asr.2023.03.036 · Cited by 125 · see citing works
Self-citationPauly, L., Rharbaoui, W., Shneider, C., Rathinam, A., Gaudillière, V., & Aouada, D. (2023) · Acta Astronautica · doi:10.1016/j.actaastro.2023.08.001 · Cited by 121 · see citing works
Deep learning-based spacecraft relative navigation methods: A survey
Song, J., Rondao, D., & Aouf, N. (2022) · Acta Astronautica · doi:10.1016/j.actaastro.2021.10.025 · Cited by 106 · see citing works
Satellite Pose Estimation Competition 2021: Results and Analyses
Park, T. H., Märtens, M., Jawaid, M., Wang, Z., Chen, B., Chin, T.-J., Izzo, D., & D’Amico, S. (2023) · Acta Astronautica · doi:10.1016/j.actaastro.2023.01.002 · Cited by 91 · see citing works
Self-citationPark, T. H., & D’Amico, S. (2023) · Journal of Guidance, Control, and Dynamics · doi:10.2514/1.g007387 · Cited by 63 · see citing works
Self-citationCosmas, K., & Kenichi, A. (2020) · Aerospace · doi:10.3390/aerospace7110159 · Cited by 51 · see citing works
Huo, Y., Li, Z., & Zhang, F. (2020) · IEEE Access · doi:10.1109/access.2020.3041415 · Cited by 44 · see citing works
PVSPE: A pyramid vision multitask transformer network for spacecraft pose estimation
Yang, H., Xiao, X., Yao, M., Xiong, Y., Cui, H., & Fu, Y. (2024) · Advances in Space Research · doi:10.1016/j.asr.2024.05.011 · Cited by 41 · see citing works
Deep learning for real-time satellite pose estimation on tensor processing units
Lotti, A., Modenini, D., Tortora, P., Saponara, M., & Perino, M. A. (2023) · Journal of Spacecraft and Rockets · doi:10.2514/1.a35496 · Cited by 39 · see citing works
Beierle, C., & D’Amico, S. (2019) · Journal of Spacecraft and Rockets · doi:10.2514/1.a34337 · Cited by 29 · see citing works
Self-citationTowards Bridging the Space Domain Gap for Satellite Pose Estimation using Event Sensing
Jawaid, M., Elms, E., Latif, Y., & Chin, T.-J. (2023) · 2023 IEEE International Conference on Robotics and Automation (ICRA) · doi:10.1109/icra48891.2023.10160531 · Cited by 32 · see citing works
Mobile-URSONet: an embeddable neural network for onboard spacecraft pose estimation
Posso, J., Bois, G., & Savaria, Y. (2022) · 2022 IEEE International Symposium on Circuits and Systems (ISCAS) · doi:10.1109/iscas48785.2022.9937721 · Cited by 27 · see citing works
6D Object Pose Estimation from Approximate 3D Models for Orbital Robotics
Ulmer, M., Durner, M., Sundermeyer, M., Stoiber, M., & Triebel, R. (2023) · 2023 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) · doi:10.1109/iros55552.2023.10341511 · Cited by 23 · see citing works
ChiNet: Deep recurrent convolutional learning for multimodal spacecraft pose estimation
Rondao, D., Aouf, N., & Richardson, M. A. (2022) · IEEE Transactions on Aerospace and Electronic Systems · doi:10.1109/taes.2022.3193085 · Cited by 24 · see citing works
Pérez-Villar, J. I. B., García-Martín, Á., Bescós, J., & Escudero-Viñolo, M. (2024) · IEEE Transactions on Aerospace and Electronic Systems · doi:10.1109/taes.2023.3306731 · Cited by 22 · see citing works
Pose Estimation for Non-Cooperative Rendezvous Using Neural Networks
Sharma, S., & D'Amico, S. (2019) · arXiv · Cited by 40 · see citing works
Self-citationSpacecraft Homography Pose Estimation with Single-Stage Deep Convolutional Neural Network
Chen, S., Yang, W., Wang, W., Mai, J., Liang, J., & Zhang, X. (2024) · Sensors · doi:10.3390/s24061828 · Cited by 12 · see citing works
Park, T. H., Bosse, J., & D'Amico, S. (2021) · arXiv · doi:10.48550/arxiv.2108.05529 · Cited by 10 · see citing works
Self-citationNon-Cooperative Spacecraft Pose Estimation Based on Feature Point Distribution Selection Learning
Yuan, H., Chen, H., Wu, J., & Kang, G. (2024) · Aerospace · doi:10.3390/aerospace11070526 · Cited by 9 · see citing works
Investigating vision transformers for bridging domain gap in satellite pose estimation
Lotti, A., Modenini, D., & Tortora, P. (2023) · Studies in Computational Intelligence · doi:10.1007/978-3-031-25755-1_20 · Cited by 7 · see citing works
Evaluating the Reliability of Vision Transformers for Space Robotics Applications
Bodmann, P. R., Rech, P., & Saveriano, M. (2024) · 2024 International Conference on Space Robotics (iSpaRo) · doi:10.1109/isparo60631.2024.10687576 · Cited by 5 · see citing works
FastPose-ViT: A Vision Transformer for Real-Time Spacecraft Pose Estimation
Ancey, P., Price, A., Javed, S., & Salzmann, M. (2026) · 2026 IEEE/CVF Winter Conference on Applications of Computer Vision (WACV) · doi:10.1109/wacv61042.2026.00760
- Authors
- Sharma, S., Beierle, C., & D'Amico, S.
- Affiliations
- Stanford University (US)
- ORCID
- Sharma, Sumant
- Venue
- 2018 IEEE Aerospace Conference
- Publisher
- IEEE
- Published
- 2018-03
- Pages
- 1-12
- ISSN
- 1095-323X (ISSN-L)
- OpenAlex
- W2810225457
- Cited by
- 207
- References
- 19
- Citations by year
- 2026: 21 · 2025: 27 · 2024: 39 · 2023: 30 · 2022: 34 · 2021: 25 · 2020: 24 · 2019: 7
- FWCI
- 24.30
- Type
- conference paper
- Language
- en
- Open access
- green
- arxiv.org
- Locations
- IEEE Aerospace Conference (AERO) (publishedVersion)
- arXiv (Cornell University) (submittedVersion)
- arXiv (Cornell University)
- Funders
- Deutsches Zentrum für Luft- und Raumfahrt
- King Abdulaziz City for Science and Technology
- Danmarks Tekniske Universitet
- Topics
- Physical Sciences › Engineering › Aerospace Engineering › Space Satellite Systems and Control
- Physical Sciences › Engineering › Aerospace Engineering › Spacecraft Dynamics and Control
- Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Advanced Neural Network Applications
- Keywords
- spacecraft pose estimation, monocular vision, convolutional neural networks, on-orbit servicing, synthetic image generation, illumination robustness
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
