“Artificial Intelligence based 6D Pose Estimation of Uncooperative Targets from Monocular Images ” Proceedings of the 73rd International Astronautical Congress 2022, IAC-2022, Paris, France. Motivation : Proximity operations around non-cooperative targets. Used Convolutional Neural Networks (CNNs) to find the pose of space targets from monocular images. Generated a labelled dataset of 1000 images of the Nauka spacecraft, taken from various angles and lighting conditions. Trained on Real and Synthetic Dataset. Synthetic dataset is generated from Blender (3D modelling tool) using the 3D model of Nauka spacecraft.
