Deep learning-driven adaptive optics for laser wavefront correction

2025/09/12 by Jikai Wang, Wang, Jikai, Sven Burckhard +11
Engineering · Medicine · Physics and Astronomy · #Adaptive optics and wavefront sensing #FOS: Physical sciences #Ocular and Laser Science Research #Optical Systems and Laser Technology #Optics (physics.optics)

paper · pdf · doi:10.48550/arxiv.2509.10662

openalex publication_date 2025/09/12 · openalex created_date 2025/10/12 · openalex updated_date 2026/07/28

Abstract

We report on an intensity-only and deep-learning based method for laser beam characterization that allows to predict the underlying optical field within milliseconds. A simple near-field / far-field camera setup enables online control of an adaptive optics to optimize beam quality. The robustness and precision of the method is enhanced by applying the concept of phase diversity based on spiral phase plates.

Citations

• Image-to-Image Translation with Conditional Adversarial Networks
• U-Net: Convolutional Networks for Biomedical Image Segmentation
• Zernike polynomials and atmospheric turbulence*

Related

• Wavefront Estimation From a Single Measurement: Uniqueness and Algorithms
• Image-based wavefront correction using model-free Reinforcement Learning
• Tempestas ex machina: A review of machine learning methods for wavefront control
• Using Graph Theory and a Plenoptic Sensor to Recognize Phase Distortions of a Laser Beam
• Direct Zernike Coefficient Prediction from Point Spread Functions and Extended Images using Deep Learning