Satellite Pose Estimation

Satellite pose estimation focuses on accurately determining a satellite's orientation and position in space using visual data, primarily to enable autonomous operations like docking and debris removal. Current research emphasizes robust methods that handle challenging lighting conditions and limited real-world data, often employing deep learning architectures like ResNet variations and transformers, combined with classical computer vision techniques and Kalman filters for improved accuracy and real-time performance on resource-constrained platforms like embedded SoCs. This field is crucial for advancing space situational awareness, autonomous satellite servicing, and mitigating the growing problem of space debris, with ongoing efforts to bridge the gap between simulated and real-world data through self-supervised learning and physics-informed neural networks.