Binary-star intensity ratios by the fork algorithm

Abstract
The fork algorithm for accurately estimating the intensity ratio of binary stars from speckle interferometry data is presented. For brighter stars simulation results suggest that the fork algorithm can attain a signal-to-noise ratio roughly 10 times greater than that of other algorithms, such as triple correlation and shift-and-add. The results of the application of the algorithm to Capella (Alpha Aurigae) data are described.