Comparison of Phase-Unwrapping Methods for Adaptive Optics Wavefront Sensing

TL;DR

This paper compares four phase unwrapping methods for adaptive optics wavefront sensing, analyzing their performance in real-time atmospheric turbulence correction scenarios.

Contribution

It provides a systematic evaluation of four mature phase unwrapping algorithms under realistic telescope boundary conditions.

Findings

Fast2D performs best in open aperture conditions.

Zernike method is robust near aperture edges.

DFT and LSPV show limitations with boundary effects.

Abstract

Advanced wavefront sensors (WFS) are essential for enabling new science cases for telescopes that utilize adaptive optics (AO) systems. While complex field WFS -- those that estimate the electric field phase and amplitude through interference or diffraction effects -- can achieve extraordinary sensitivity compared to existing devices, they typically reconstruct the wrapped phase of the measured wavefront, which must then be unwrapped for correction by continuous-surface deformable mirrors (DM). Another requirement is that the phase function must be unwrapped within 1 millisecond or faster for real-time AO operations. Using simulations of atmospheric turbulence that follow a Kolmogorov spectrum, we study four prevalent and mature phase unwrapping methods: Fast2D, Zernike Gradient (Zernike), Discrete Fourier Transform (DFT), and Least Squares Principle Value (LSPV). In this paper, we examine the strengths and limitations of each method. In particular, we consider performance with and without a binary circular aperture boundary that defines the edge of monolithic telescopes.