Spacecraft Pose Estimation Dataset (SPEED)
收藏资源简介:
Spacecraft Pose Estimation Dataset (SPEED) 是由欧洲空间局和斯坦福大学的空间会合实验室共同创建的,首个公开可用的机器学习数据集,用于卫星姿态估计。该数据集包含近15000张合成图像和300张真实图像,这些图像来源于Tango卫星的PRISMA任务。数据集的创建过程涉及使用光学仿真器软件生成合成图像,并在TRON设施中捕捉真实图像。SPEED数据集主要用于卫星姿态估计算法的训练和测试,特别是在自主轨道服务和碎片移除任务中,解决缺乏空间图像训练和验证算法的问题。
The Spacecraft Pose Estimation Dataset (SPEED) was jointly created by the European Space Agency and the Space Rendezvous Laboratory of Stanford University, and it is the first publicly available machine learning dataset dedicated to satellite pose estimation. The dataset contains approximately 15,000 synthetic images and 300 real-world images, all sourced from the Tango satellite of the PRISMA mission. The development of the SPEED dataset involved generating synthetic images using optical simulator software and capturing real-world imagery at the TRON facility. The SPEED dataset is primarily used for training and testing satellite pose estimation algorithms, particularly in autonomous orbital servicing and space debris removal missions, to address the shortage of space imagery for training and validating relevant algorithms.

- 1Satellite Pose Estimation Challenge: Dataset, Competition Design and Results欧洲空间局 · 2020年



