Correlation-based Shack-Hartmann wavefront slope sensing in strong turbulence

The Shack-Hartmann wavefront sensor is the most common wavefront sensor used today in adaptive optics. While the most common way to employ a Shack-Hartmann wavefront sensor is with a centroiding-based algorithm, recent studies have shown how its accuracy can be increased by using a correlation-based algorithm. This paper compares the accuracy of a correlation-based Shack-Hartmann wavefront sensor to a centroiding-based Shack-Hartmann wavefront sensor by measuring the distortions of a point source viewed through strong turbulence where scintillation, variations in the amplitude of the field, plays an important role. In particular, this paper demonstrates that a correlation-based algorithm generally does outperform the centroiding-based algorithm, but by using a simple threshold a centroiding-based Shack-Hartmann can produce similar results as a correlation-based Shack-Hartmann. Also, in very low signal levels, such as when the average signal level is below the detector read noise, a regular centroiding-based Shack-Hartmann outperforms all other methods tested.