Navigating spacecraft and mapping the surface in the vicinity of small Solar System bodies is historically challenging. While surface landmarks and related optical measurements are key to enhance navigational performance in such an environment, processing close-proximity images relies on ground operations and requires demanding, human-led coordination between orbit determination, shape modeling, and image processing tasks. In this paper, we present a visiononly autonomous pipeline for terrain-relative positioning, landmark matching, and dense mapping to perform such optical-navigation tasks onboard. We leverage stereophotogrammetric techniques, in particular structure-from-motion for camera pose estimation and dense stereo matching for shape reconstruction. The pipeline requires no prior knowledge of the environment or the observer’s state. Further, we propose a novel feature matching algorithm based on geometric invariant properties between landmarks, rather than local image data, for feature description; we show its robustness to lighting and viewpoint variations. We assess performance for each pipeline phase using real imagery from the OSIRIS-REx mission, comparing our results with the mission’s reconstructed estimates. We demonstrate good performance of the proposed approach using close-proximity images of the rubble-pile asteroid Bennu, where surface shadowing and viewpoint changes are prominent.
AUTONOMOUS NAVIGATION AND DENSE SHAPE RECONSTRUCTION USING STEREOPHOTOGRAMMETRY AT SMALL CELESTIAL BODIES
Proceedings of the 44th Annual American Astronautical Society Guidance, Navigation, and Control Conference, 2022 ; Kapitel : 73 ; 1325-1347
2024-01-01
23 pages
Aufsatz/Kapitel (Buch)
Elektronische Ressource
Englisch
Autonomous Navigation Using Celestial Objects
NTRS | 1999
|AUTONOMOUS ONLINE NAVIGATION AND PARAMETER ESTIMATION IN THE VICINITY OF SMALL CELESTIAL BODIES
Europäisches Patentamt | 2024
|British Library Conference Proceedings | 2012
|CELESTIAL NAVIGATION SYSTEM FOR AN AUTONOMOUS VEHICLE
Europäisches Patentamt | 2017
|