Dynamic-Spacecraft Pose Estimation Dataset (D-SPEED)
Bibliographic record
Abstract
D-SPEED: A Synthetic Benchmark for Temporal Spacecraft Pose Estimation D-SPEED is a synthetic dataset designed for deep learning-based relative pose estimation of non-cooperative spacecraft, from both still images and video. It extends prior datasets (SPEED, SPEED+) by introducing temporally coherent video sequences, detailed motion metadata, and high-resolution rendering using Unreal Engine 5. It contains: 60,000 high-resolution still images of the Tango spacecraft under varied poses and lighting. 21 video sequences (25 FPS, 1500 frames each) covering 11 distinct motion trajectories (e.g., docking, formation flying, inspection), with per-frame ground-truth 6-DoF poses. Camera intrinsics and predefined train/val/test splits to support reproducible training and evaluation. Compared to previous datasets, D-SPEED enables the development of temporal pose estimation algorithms through continuous video streams and annotated motion events (e.g., accelerations, camera/satellite motion). 🎥 Teaser video: https://youtu.be/AbIYOj8LuNY 🛠 Companion tools for 2D keypoint and bounding box generation, visualization, ground-truth generation, and basic evaluation workflows are available in the open-source repository:👉 https://github.com/possoj/Spacecraft-Pose-Estimation-Framework 📄 For trajectory metadata, sampling distributions, and details of the video sequence generation process, please refer to the associated paper (under review at IEEE Transactions on Aerospace and Electronic Systems) and the associated PhD thesis. 🎮 Rendering performed by Rexys using their Unreal Engine-based toolchain: https://rexys.io 📝 If you use this dataset, please cite: [1] Julien Posso, Guy Bois, and Yvon Savaria, Dynamic-Spacecraft Pose Estimation Dataset (D-SPEED), Zenodo, 2025. https://doi.org/10.5281/zenodo.15851302 [2] Julien Posso, Estimation de pose de véhicules spatiaux non coopératifs à partir de réseaux de neurones – De l'image monoculaire à l'implémentation embarquée temps réel et à l'analyse temporelle, PhD thesis, Polytechnique Montréal, 2025. https://publications.polymtl.ca/67849/
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.001 | 0.002 |
| Meta-epidemiology (narrow) | 0.004 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.002 | 0.002 |
| Science and technology studies | 0.001 | 0.001 |
| Scholarly communication | 0.001 | 0.001 |
| Open science | 0.002 | 0.001 |
| Research integrity | 0.002 | 0.002 |
| Insufficient payload (model declined to judge) | 0.010 | 0.014 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".