CENTROIDING AND SIZING OPTIMIZATION OF ELLIPSOID IMAGE PROCESSING USING NONLINEAR-LEAST SQUARES

2018 Univelt Inc. All rights reserved. This paper covers the development and application of image processing algorithms used on images of celestial target bodies, specifically the Earth and Moon, when performing optical navigation (OpNav). These algorithms were originally developed by Texas A&M under contract with NASA Johnson Space Center as a backup navigation solution in the case of a loss of communications with ground stations on Earth. The algorithms work by first pre-processing the image ensuring that it can be processed. Then, an initial estimate of the apparent size and centroid of the observed body is produced. This initial estimate is then refined by using nonlinear least squares to fit a functional description to the limb of the observed body. The focus of this paper will be to provide the mathematical development of the nonlinear least squares and also demonstrate its implementation in the Orion OpNav system. Results of both real image and synthetic image processing will be presented and recommendations for further algorithm implementation will be presented.

CENTROIDING AND SIZING OPTIMIZATION OF ELLIPSOID IMAGE PROCESSING USING NONLINEAR-LEAST SQUARES Conference Paper