Real-Time, Flight-Ready, Non-Cooperative Spacecraft Pose Estimation Using Monocular Imagery

TL;DR

This paper introduces a CNN-based monocular pose estimation system for spacecraft that is accurate, efficient, and capable of real-time operation on low-power hardware, using synthetic data for training and generalizing to real space imagery.

Contribution

A novel CNN approach for monocular spacecraft pose estimation that combines synthetic data training with real-world generalization and real-time performance on flight hardware.

Findings

Achieves state-of-the-art accuracy in pose estimation

Successfully generalizes from synthetic to real space imagery

Operates in real-time on low-power hardware

Abstract

A key requirement for autonomous on-orbit proximity operations is the estimation of a target spacecraft's relative pose (position and orientation). It is desirable to employ monocular cameras for this problem due to their low cost, weight, and power requirements. This work presents a novel convolutional neural network (CNN)-based monocular pose estimation system that achieves state-of-the-art accuracy with low computational demand. In combination with a Blender-based synthetic data generation scheme, the system demonstrates the ability to generalize from purely synthetic training data to real in-space imagery of the Northrop Grumman Enhanced Cygnus spacecraft. Additionally, the system achieves real-time performance on low-power flight-like hardware.