Spacecraft Pose Estimation Dataset (SPEED)

The SPEED dataset is the official dataset of ESA's Kelvins "Pose Estimation challenge" in collaboration with Stanford Universitiy's Space Rendezvous Lab (SLAB). It features images and poses of the Tango spacecraft (PRISMA mission), 12000 of them generated by SLAB's Optical Simulator using a high fidelity texture model and 300 images from the TRON facility, using a physical mock-up model of Tango. The goal of the competition was estimate the relative pose (distance and orientation) from pixel images only. Detailed information about the original competition can be found at https://kelvins.esa.int/satellite-pose-estimation-challenge/ A follow-up competition with a larger and improved dataset (SPEED+) is available on Zenodo as well: https://zenodo.org/record/5588480 A publication about the results of the pose estimation challenge has been published as Kisantal, Mate, et al. "Satellite pose estimation challenge: Dataset, competition design, and results." IEEE Transactions on Aerospace and Electronic Systems 56.5 (2020): 4083-4098.

SPEED数据集是欧洲空间局（ESA）与斯坦福大学空间交会实验室（SLAB）联合推出的Kelvins“姿态估计挑战赛”官方数据集。该数据集收录了PRISMA任务中探戈号航天器的图像与姿态数据：其中12000组数据由SLAB的光学模拟器基于高保真纹理模型生成，另有300组图像源自TRON设施，该设施采用了探戈号航天器的实体样机模型。本次挑战赛的核心目标为仅通过像素级图像，估算航天器间的相对姿态（包括相对距离与方位）。有关原始挑战赛的详细信息，可访问：https://kelvins.esa.int/satellite-pose-estimation-challenge/。此外，针对该挑战赛打造的升级版数据集（SPEED+）规模更大、优化程度更高，其相关后续挑战赛同样可在Zenodo平台获取：https://zenodo.org/record/5588480。有关本次姿态估计挑战赛结果的研究论文已正式发表：Kisantal, Mate 等. 卫星姿态估计挑战赛：数据集、竞赛设计与研究成果[J]. IEEE航空与电子系统汇刊, 2020, 56(5): 4083-4098.

搜集汇总

背景与挑战

背景概述

SPEED数据集是欧洲空间局与斯坦福大学合作推出的官方数据集，专门用于航天器姿态估计挑战。它包含Tango航天器的图像和姿态数据，其中12000张为高保真纹理模型生成的模拟图像，300张为物理模型拍摄的真实图像，旨在通过计算机视觉技术估计航天器的相对位置和方向。该数据集支持机器学习和计算机视觉研究，并已发布改进版本SPEED+和相关学术论文。