ChiNet: Deep Recurrent Convolutional Learning for Multimodal Spacecraft Pose Estimation
Open accessThe DOI matches its Crossref record; 65 of the 65 references it cites were found in a registry or an index.
Abstract
This paper presents an innovative deep learning pipeline which estimates the relative pose of a spacecraft by incorporating the temporal information from a rendezvous sequence. It leverages the performance of long short-term memory (LSTM) units in modelling sequences of data for the processing of features extracted by a convolutional neural network (CNN) backbone. Three distinct training strategies, which follow a coarse-to-fine funnelled approach, are combined to facilitate feature learning and improve end-to-end pose estimation by regression. The capability of CNNs to autonomously ascertain feature representations from images is exploited to fuse thermal infrared data with electro-optical red-green-blue (RGB) inputs, thus mitigating the effects of artifacts from imaging space objects in the visible wavelength. Each contribution of the proposed framework, dubbed ChiNet, is demonstrated on a synthetic dataset, and the complete pipeline is validated on experimental data.
Overview
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
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?
The cited source says what the claim attributes to it.
In the citing paper
“In this direction, recently proposed ChiNet [188] have used Long Short-Term Memory (LSTM) [194] units in modelling sequences of data for estimating the spacecraft pose.”
From A survey on deep learning-based monocular spacecraft pose estimation: Current state, limitations and prospects (2023) · cites this paper as [188]
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
Citations
Cites
Deep Residual Learning for Image Recognition
He, K., Zhang, X., Ren, S., & Sun, J. (2016) · 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2016.90 · Cited by 235,474
Hochreiter, S., & Schmidhuber, J. (1997) · Neural Computation · doi:10.1162/neco.1997.9.8.1735 · Cited by 102,318 · cited 2×
“ChiNet’s recurrent feature postprocessing module is based on the LSTM architecture [41].”
MethodAdam: A Method for Stochastic Optimization
Kingma, D. P., & Ba, J. (2014) · arXiv · doi:10.48550/arxiv.1412.6980 · Cited by 98,946
ImageNet: A large-scale hierarchical image database
Deng, J., Dong, W., Socher, R., Li, L.-J., Li, K., & Fei-Fei, L. (2009) · 2009 IEEE Conference on Computer Vision and Pattern Recognition · doi:10.1109/cvpr.2009.5206848 · Cited by 65,333 · cited 1×
“A subset of 1000 object categories of ImageNet [53] is the typical go-to choice for pre-trained networks.”
BackgroundDistinctive Image Features from Scale-Invariant Keypoints
Lowe, D. G. (2004) · International Journal of Computer Vision · doi:10.1023/b:visi.0000029664.99615.94 · Cited by 55,589 · cited 2×
“, SIFT [19], SURF [20], or the more modern ORB [21]) to obtain the matches.”
BackgroundRethinking the Inception Architecture for Computer Vision
Szegedy, C., Vanhoucke, V., Ioffe, S., Shlens, J., & Wojna, Z. (2016) · 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2016.308 · Cited by 32,138
Fischler, M. A., & Bolles, R. C. (1981) · Communications of the ACM · doi:10.1145/358669.358692 · Cited by 25,893 · cited 4×
“As the target is generally known beforehand, the followed strategies often choose to solve the model-to-image registration problem, under which the pose is retrieved via methods in [4] and [5] from correspondences between 2-D image features and 3-D model points.”
MethodInfluentialYOLO9000: Better, Faster, Stronger
Redmon, J., & Farhadi, A. (2017) · 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2017.690 · Cited by 19,892 · cited 1×
“To mitigate this, ChiNet adopts the Darknet-19 architecture (backbone of the YOLO object detector [45]), with some modifications (see Fig.”
MethodYOLOv3: An Incremental Improvement
Redmon, J., & Farhadi, A. (2018) · arXiv · doi:10.48550/arxiv.1804.02767 · Cited by 17,141 · cited 1×
“Second the network is modernised (bringing it closer to Darknet-53 [47]) by replacing all max pooling layers with a stride of 2 in the preceding convolution.”
MethodSURF: Speeded Up Robust Features
Bay, H., Tuytelaars, T., & Van Gool, L. (2006) · Lecture Notes in Computer Science · doi:10.1007/11744023_32 · Cited by 14,955 · cited 1×
“An alternative approach constructs a database by discretising the 3D object into 2D keyframes representing multiple viewpoints [18], and then using local keypoint detectors and descriptors (e.g. SIFT [19], SURF [20], or the more modern ORB [21]) to obtain the matches.”
Method
Research works printed but not linked: 1
- [23] Gansmann, M., Mongrard, O., and Ankersen, F., '3D ModelBased Relative Pose Estimation for Rendezvous and Docking Using Edge Features,' in 10 th International ESA Conference on Guidance, Navigation and Control Systems , Salzburg, Austria: ESA, 2017. -
Cited by
Pauly, 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
“In this direction, recently proposed ChiNet [137] have used Long Short-Term Memory (LSTM) [50] units in modelling sequences of data for estimating the spacecraft pose.”
BackgroundMethodGolroudbari, A. A., & Sabour, M. H. (2023) · arXiv (Cornell University) · doi:10.48550/arxiv.2302.11089 · Cited by 37
“[151] Supervised YOLO Synthetic Fault Detection [152] Supervised YOLO + MobileNet SPEED Pose Estimation [153] Supervised Faster R-CNN + HRNet SPEED Pose Estimation [154] Reinforcement Learning Deep Reinforcement Learning ✗ Landing [146] Reinforcement Learning Deep Reinforcement Learning ✗ Decision Making [155] Reinfor…”
MethodRTM-UAVDet: A Real-Time Multimodal UAV Detector
Wang, G., Jiang, Q., Jin, X., Woźniak, M., Yao, P. W. S., & Yao, S. (2024) · IEEE Transactions on Aerospace and Electronic Systems · doi:10.1109/taes.2024.3446460 · Cited by 33
TransFusionOdom: Transformer-Based LiDAR-Inertial Fusion Odometry Estimation
Sun, L., Ding, G., Qiu, Y., Yoshiyasu, Y., & Kanehiro, F. (2023) · IEEE Sensors Journal · doi:10.1109/jsen.2023.3302401 · Cited by 30
Golroudbari, A. A., & Sabour, M. H. (2023) · Authorea · doi:10.22541/au.168664884.43899660/v1 · Cited by 25
Zhang, Y., You, L., Yang, L., Zhang, Z., Nie, X., & Zhang, B. (2024) · IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW) · doi:10.1109/cvprw63382.2024.00678 · Cited by 22
“In [39], the ChiNet method is proposed by incorporating Long Short-Term Memory (LSTM) units [15] for modeling sequences of data in spacecraft pose estimation.”
MethodDeep-Learning-Based Direct Attitude Estimation for Uncooperative Known Space Objects
Qiao, S., Zhang, H., Xie, F., & Jiang, Z. (2023) · IEEE Transactions on Aerospace and Electronic Systems · doi:10.1109/taes.2023.3325801 · Cited by 19
Robust deep learning LiDAR-based pose estimation for autonomous space landers
Chekakta, Z., Zenati, A., Aouf, N., & Dubois-Matra, O. (2022) · Acta Astronautica · doi:10.1016/j.actaastro.2022.08.049 · Cited by 19
Self-citationAI-based monocular pose estimation for autonomous space refuelling
Rondao, D., He, L., & Aouf, N. (2024) · Acta Astronautica · doi:10.1016/j.actaastro.2024.04.003 · Cited by 14
Self-citationDomain Generalization for In-Orbit 6D Pose Estimation
Legrand, A., Detry, R., & De Vleeschouwer, C. (2025) · Journal of Aerospace Information Systems · doi:10.2514/1.i011498 · Cited by 10
“LIDAR systems can produce 3D point clouds [8].”
Relative Pose Determination for Noncooperative Spacecraft Under Noninertial Observation Frame
Chen, H., Shao, X., Yang, H., Hu, Q., & Li, D. (2025) · IEEE Transactions on Aerospace and Electronic Systems · doi:10.1109/taes.2025.3562539 · Cited by 9
Golroudbari, A. A., & Sabour, M. H. (2023) · Authorea · doi:10.22541/au.168664884.43899660/v2 · Cited by 8
Hybrid deep learning based monocular pose estimation for autonomous space docking operations
Khalil, S., Wang, Z., & Aouf, N. (2025) · Acta Astronautica · doi:10.1016/j.actaastro.2025.10.010 · Cited by 7
Self-citationWang, Z., & Aouf, N. (2024) · Acta Astronautica · doi:10.1016/j.actaastro.2024.07.052 · Cited by 6
Self-citationWang, Z., Aouf, N., Pizarro, J., & Honvault, C. (2024) · Advances in Space Research · doi:10.1016/j.asr.2024.11.054 · Cited by 5
Self-citationMotion Aware ViT-based Framework for Monocular 6-DoF Spacecraft Pose EstimationPreprint
Sosa, J., Pineau, D., Rathinam, A., Shabayek, A., & Aouada, D. (2025) · arXiv (Cornell University) · doi:10.48550/arxiv.2509.06000 · Cited by 5
“For example, ChiNet [3] incorporates a CNN-LSTM architecture to regress the 6-DoF pose from image sequences, following an end-to-end strategy.”
Deep recurrent neural network-based satellite indirect pose tracking with adaptive Huber loss
Chen, Z., Li, Q., Zhong, R., & Gui, H. (2025) · Advances in Space Research · doi:10.1016/j.asr.2025.11.021 · Cited by 4
A review of fine-grained object perception in extreme space environments
ZHANG, C., Jiaqi, Y. A. N. G., Yan, Q., Qi, Z., Yang, Z. O. U., & Zhang, Y. (2025) · Scientia Sinica Informationis · doi:10.1360/ssi-2024-0394 · Cited by 5
An End-to-End Pose Estimation Network for Multiscale Space Non-cooperative Objects
Cao, Y., & Cheng, X. (2023) · International Conference on Control and Robotics Engineering (ICCRE) · doi:10.1109/iccre57112.2023.10155575 · Cited by 4
“At present, the methods of monocular pose estimation of space non-cooperative objects using convolution neural network (CNN) are mainly divided into two categories, one is based on three-stage network, and the other is based on end-toend network [4].”
MethodVIPE: Visible and Infrared Fused Pose Estimation Framework for Space Noncooperative Objects
Zhang, Z., Zhou, D. X., Hu, Y., Ma, W., Sun, G., & Zhang, Y. (2025) · Sensors · doi:10.3390/s25216664 · Cited by 3
“Rondao et al. [43] developed the ChiNet framework, which effectively enhances RGB images by incorporating long-wave infrared data, thereby providing richer feature representations while mitigating the impact of artifacts commonly encountered in visible-light imaging of spatial objects.”
NeRF-based Visualization of 3D Cues Supporting Data-Driven Spacecraft Pose Estimation
Legrand, A., Detry, R. J., & Vleeschouwer, C. D. (2025) · International Conference on Space Robotics (iSpaRo) · doi:10.1109/isparo66239.2025.11436169 · Cited by 3
“Although several ones have been considered in the literature [4], such as LIDAR systems [5], infrared [6], [7], time-of-flight [8] or stereo [9] cameras, many recent works focused on the use of monocular cameras due to their lower cost, mass, power consumption, or bulkiness [3].”
Chekakta, Z., & Aouf, N. (2024) · IEEE Sensors Journal · doi:10.1109/jsen.2024.3435748 · Cited by 3
“Hence, in the recent development of rendezvous visual-based pose estimation approaches, the IPIP step has been completely shifted to a deep learning model able to generalize for different image conditions [5].”
BackgroundMethodInfluentialSelf-citationYuan, H., Wu, J.-F., & Kang, G. (2025) · Aerospace Science and Technology · doi:10.1016/j.ast.2025.111240 · Cited by 2
Noncooperative Spacecraft Pose Measurement Without Prior Knowledge Based on SAM2
Guo, P., Jiang, C., Long, C., Hu, Q., & Li, D. (2026) · IEEE Transactions on Instrumentation and Measurement · doi:10.1109/tim.2026.3654753 · Cited by 2
“These studies encompass both dataset development [10] and intelligent model design [11], [12], [13], [14], [15], [16], [17], [18], [19].”
Motion Posture Detection Algorithm and Application based on Computer Vision and DNN
Meng, H., & Shi, S. (2025) · International Conference on Mobile Networks and Wireless Communications (ICMNWC) · doi:10.1109/icmnwc66779.2025.11354255 · Cited by 1
“(cid:19) (cid:20) By stacking multiple layers of ST-GCN, the network can gradually abstract high-order spatiotemporal features from the original joint coordinates [13].”
Koushafar, M., & Kosari, A. (2026) · Acta Astronautica · doi:10.1016/j.actaastro.2026.07.027 · Cited by 1
Golroudbari, A. A., & Sabour, M. H. (2023) · Authorea · doi:10.22541/au.168727976.63672670/v1 · Cited by 1
SpY: A Context-Based Approach to Spacecraft Component DetectionPreprint
Mahendrakar, T., White, R. T., & Tiwari, M. (2024) · arXiv (Cornell University) · doi:10.48550/arxiv.2406.18709 · Cited by 1
“Notable are participating works [15], [16], [17], [18], [19] in ESA’s spacecraft pose estimation challenges for non-cooperative spacecraft based on the SPEED [20] and SPEED+ [21] datasets.”
BackgroundTALON: Token-Aligned Lightweight Adapters for 6-DoF Spacecraft Pose EstimationPreprint
Ali, A., Chenni, N. E., Rathinam, A., & Aouada, D. (2026) · arXiv (Cornell University) · doi:10.48550/arxiv.2605.31217 · Cited by 1
“The few exceptions incorporate temporal information through full fine-tuning of recurrent architectures [18, 32] or auxiliary optical flow networks [35, 41].”
Artificial Intelligence-Enabled Space Robot Operations: Technologies, Challenges and ProspectsPreprint
Huang, Z., Chen, G., Hao, Z., Tang, G., Zong, J., Ban, G., Wang, J., Lv, H., Ren, C., & Liu, S. (2026) · arXiv (Cornell University) · doi:10.48550/arxiv.2609.16880 · Cited by 1
“Rondao et al. proposed ChiNet [72], which integrates RGB, thermal-infrared, and temporal features to improve the continuity of relative pose estimation under complex imaging conditions.”
HAT: Hypothesis-Anchored Tracking for Video Monocular Spacecraft Pose EstimationPreprint
Lopo, A., Dehban, A., & Ventura, R. (2026) · arXiv (Cornell University) · doi:10.48550/arxiv.2609.21597 · Cited by 1
“Prior work uses navigation filters [10], sequence networks [11], and cross-domain tracking [8].”
Relative Pose Determination of a Non-Cooperative Spacecraft using Thermal Imagery
Budhkar, A. (2024) · Carleton University's Institutional Repository (MacOdrum Library, Carleton University) · doi:10.22215/etd/2024-16188
Trend Shift in Pose Estimation: Toward Continuous Representation
Bang, S., Lee, C., Ryu, J., Lee, H. T., & Paek, J. (2024) · International Conference on Information and Communication Technology Convergence (ICTC) · doi:10.1109/ictc62082.2024.10827211
Raja, N. K., Ramesh, J. V. N., El–Ebiary, Y. A. B., Muniyandy, E., Reddy, N. K., Kumar, V. R., & Devarasetty, P. (2025) · International Journal of Advanced Computer Science and Applications · doi:10.14569/ijacsa.2025.0160463
“This involves training a deep network via synthetic images and testing it on genuine information, the latter of which are usually hard to come by before the mission begins, however the earlier kind might be produced in huge quantities [8].”
基于视觉与深度学习的非合作空间目标识别与位姿估计方法研究进展
Jianxiang, J., & Yiquan, W. (2025) · Journal of Image and Graphics · doi:10.11834/jig.250435
Bechis, L., Sarvadon, J.-L., Ricioppo, P., & Mancini, M. (2026) · Control Engineering Practice · doi:10.1016/j.conengprac.2026.107134
“To address these issues, [22] proposed a Recurrent Convolutional Neural Network (RCNN) with monocular camera images, leveraging its temporal modeling capability to incorporate information from previous inputs for pose estimation.”
Zhao, D., Yue, Z., Zhang, W., Yuan, J., Ma, W., Ban, H., Li, S. L., & Weiwei, L. (2026) · Aerospace · doi:10.3390/aerospace13070583
“Rondão et al. adopted a CNN + LSTM temporal modeling scheme to enable sequential attitude estimation with multimodal robustness [14].”
Chen, D., Liu, D., Wang, Y., Ma, S., Fu, Z., Jiang, J., & Wang, X. (2026) · SSRN Electronic Journal · doi:10.2139/ssrn.7274279
Physics-based diffusion model for thermal image generation in space orbital rendez-vous scenarios
Guerazem, S., & Aouf, N. (2026) · Acta Astronautica · doi:10.1016/j.actaastro.2026.08.058
Self-citationLiang, Z., Liu, G., Song, J., & Li, X. (2026) · International Symposium on Aerospace Engineering and Systems (ISAES) · doi:10.1109/isaes70982.2026.11669185
“For spacecraft pose estimation, Rondao et al. [10] proposed ChiNet, a deep recurrent convolutional framework that uses RGB and thermal infrared information to mitigate artifacts in visible-wavelength imaging.”
Guerazem, S., Rondao, D., & Aouf, N. (2026) · IEEE Transactions on Aerospace and Electronic Systems · doi:10.1109/taes.2026.3731320 · Cited by 1
“To jointly optimize both tasks, we adopt the ho-moscedastic uncertainty weighting method proposed by Kendall et al. [56], also used by Rondao et al. [9].”
InfluentialSelf-citationUncertainty-aware monocular spacecraft pose estimation with covariance-weighted keypoints
Zhai, F., Zhang, C., He, Z., & Xu, D. (2026) · Aerospace Systems · doi:10.1007/s42401-026-00554-2
Çakır, A., Campbell, A., Porter, F., Savitski, V. G., Hall, I., Longino, J., Zabalza, J., Vasile, M., Murray, P., & Feng, J. (2026) · Acta Astronautica · doi:10.1016/j.actaastro.2026.10.048
Metadata
- Authors
- Rondao, D., Aouf, N., & Richardson, M. A.
- Affiliations
- City, University of London (GB)
- University of London (GB)
- Cranfield University (GB)
- Venue
- IEEE Transactions on Aerospace and Electronic Systems
- Published
- 2022
- Pages
- 1-13
- ISSN
- 0018-9251 (ISSN-L)
- 1557-9603
- 2371-9877
- OpenAlex
- W3195488172
- Cited by
- 24
- References
- 62
- Citations by year
- 2026: 3 · 2025: 5 · 2024: 8 · 2023: 7 · 2022: 1
- FWCI
- 2.34
- Type
- journal article
- Language
- en
- Open access
- green
- arxiv.org
- Licence
- Locations
- IEEE Transactions on Aerospace and Electronic Systems (publishedVersion)
- arXiv (Cornell University) (submittedVersion)
- City Research Online (City University London) (submittedVersion)
- CERES (Cranfield University) (submittedVersion, cc-by-nc)
- arXiv (Cornell University)
- Topics
- Physical Sciences › Engineering › Aerospace Engineering › Space Satellite Systems and Control
- Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Advanced Neural Network Applications
- Physical Sciences › Engineering › Aerospace Engineering › Spacecraft Dynamics and Control
- Keywords
- spacecraft pose estimation, relative pose estimation, long short-term memory, convolutional neural networks, multimodal fusion, thermal infrared imaging, RGB imaging, coarse-to-fine training, representation learning
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
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.