A comparison of digital centering algorithms

Abstract

Synthetically produced star images have been used to investigate five one-dimensional centering algorithms for a wide range of observing conditions. This work confirms the earlier photographic results of van Altena and his collaborators in finding that the Gaussian fit and median routines are good centering methods, while the moment and derivative-search routines are found to be inferior. A modification to the moment analysis that suppresses the sky background below a prechosen threshold level is found to be the best centering method when the sky-background level is significant. Detailed discussions of the advantages and disadvantages of each of the methods considered herein are given as well as a discussion of the real-time image-processing techniques used by the USNO 7 in. transit circle.