Optical Navigation Algorithm Performance

Abstract

There is a wide variety of optical navigation (OpNav) techniques that can be used to extract observables from images of natural bodies. Each of these techniques has a number of strengths and weaknesses and domains where they are most applicable. In this paper, we compare the performance of some of the most commonly used OpNav techniques across a variety of orbital regimes and a variety of body types through the use of synthetic images. Specifically, we consider the techniques of analytic model fitting, phase corrected moment estimation, limb-scanning, ellipsoid matching, and cross correlation using synthetic images of a tri-axial ellipsoid, the asteroid Bennu, and the comet 67P/Churyumov-Gerasimenko. For each technique, regime, and body, we examine the overall accuracy and the type of information available. The resulting information provides a useful tool for understanding which techniques are best suited for a given image, as well as for understanding the relative performance of each technique.