Deep learning-based spacecraft pose estimation for the pre-capture phase scenario

The increasing population of space debris in Low-Earth Orbit (LEO) poses a significant threat to operational satellites and future space endeavors. To address this challenge, leading aerospace companies worldwide are developing on-orbit servicing and debris removal satellites. These servicer satellites will be capable of complex orbital operations, such as capturing tumbling defunct spacecraft. A fundamental requirement for the success of such missions is the development of accurate spacecraft pose estimation, which provides the servicer's guidance and control system with precise information about the target spacecraft's attitude. This paper addresses the study of such a pipeline using deep learning and classical computer vision algorithms