Satellite Pose Estimation Challenge

Do you want do collaborate with researchers from Stanford and ESA on a challenging machine learning problem? Come join our Satellite Pose Estimation Challenge! Pose estimation is a key for future on-orbit servicing and debris removal missions. For more information about the competition visit our competition site: https://kelvins.esa.int/satellite-pose-estimation-challenge/

Content

This dataset contains a training set of 13000 high resolution synthetic images of a satellite with corresponding ground truth pose labels, a synthetic test set with 2998 images, and another test set of 300 real images. For more information visit our competition site.