Learning-Based Pose Estimation of Non-Cooperative Spacecrafts with Uncertainty Prediction

This is the official implementaion of paper Learning-Based Pose Estimation of Non-Cooperative Spacecrafts with Uncertainty Prediction, which is accepted in Aerospace 2022. This repository contains Pytorch training code and evaluation code.

The framework of our method.

Dataset

The dataset SPEED used in our paper can be found at https://kelvins.esa.int/satellite-pose-estimation-challenge/.

Pretrained Model

YOLO_small, YOLO_tiny, YOLO_nano

Citation

If you find our code useful, you can cite us using the following bibTex:

@article{li2022learning, title={Learning-Based Pose Estimation of Non-Cooperative Spacecrafts with Uncertainty Prediction}, author={Li, Kecen and Zhang, Haopeng and Hu, Chenyu}, journal={Aerospace}, volume={9}, number={10}, pages={592}, year={2022}, publisher={MDPI}}