A virtual test environment for validating spacecraft optical navigation

Abstract

The use of machine vision to guide robotic spacecraft is being considered for a wide range of missions, such as planetary approach and landing, asteroid and small body sampling operations and in-orbit rendezvous and docking. Numerical simulation plays an essential role in the development and testing of such systems, which in the context of vision-guidance means that realistic sequences of navigation images are required, together with knowledge of the groundtruth camera motion. Computer generated imagery (CGI) offers a variety of benefits over real images, such as availability, cost, flexibility and knowledge of the ground truth camera motion to high precision. However, standard CGI methods developed for terrestrial applications lack the realism, fidelity and performance required for engineering simulations.In this paper, we present the results of our ongoing work to develop a suitable CGI-based test environment for spacecraft vision guidance systems. We focus on the various issues involved with image simulation, including the selection of standard CGI techniques and the adaptations required for use in space applications. We also describe our approach to integration with high-fidelity end-to-end mission simulators, and summarise a variety of European Space Agency research and development projects that used our test environment.

Autonomous visual recognition of known surface landmarks for optical navigation around asteroids

“…The first stage in database construction is to generate a large number of simulated images of the asteroid under a wide range of viewing and illumination angles. In this work we use the Planet and Asteroid Natural scene Generation Utility (PANGU) tool (25) , which is a high-fidelity image simulation package tailored to space environments, and is used to test and validate spacecraft vision-based guidance systems in software.…”