High-bandwidth nonlinear curvature wavefront sensing based on dual-defocused scheme and optimal parameter selection

Abstract

In adaptive optics, nonlinear curvature wavefront sensing (nlCWFS) has emerged as an effective tool in the challenging field of wavefront sensing due to its excellent sensitivity, dynamic range, and speed. Standard nlCWFS typically utilizes four intensity patterns acquired at four different defocused distances along the optical axis as amplitude constraints in numerical iterative algorithms to reconstruct the equivalent turbulence wavefront at the pupil plane. When the initial guess is far from the true phase, the iteration becomes time-consuming and may even converge to local optima. Besides, it also involves a highly complex for pupil image acquisition, and limits the system to achieve high temporal bandwidth when acquiring real-time pupil images in weak guide star scenarios. To alleviate the aforementioned issues in nlCWFS, we propose a highly robust wavefront sensing scheme with only two defocused intensity patterns. By using the wavefront phase obtained from the previous frame, the efficiency of iteration in nlCWFS can be significantly improved for the subsequent wavefront calculation. This method is capable of maintaining excellent wavefront sensing accuracy and speed even in under-sampling conditions with only two defocused images, thereby demonstrating its ability to reduce the burden on image acquisition and its potential for weak guide star applications. Additionally, we innovatively introduce the concept of a score-map to address the inherent challenge of optimal parameter selection in nlCWFS, enabling it and other similar wavefront sensing methods to adaptively adjust parameters and to achieve scenario-specific optimal performance dynamically.