Scholarlink

<p>We propose a convolutional neural network–based algorithm for retrieving the wavefront shape of a beam using in-focus and out-of-focus intensity distributions. The neural network is trained on numerically generated data. In numerical experiments, the trained model enabled wavefront shape reconstruction with a relative error below 20% for root-mean-square (rms) wavefront distortion in the range of (0.05–0.25)λ, where λ is the radiation wavelength. In a physical experiment conducted on the PEARL laser facility, for a beam with an 18 cm aperture, equal to one-quarter of the focal length, the relative error was approximately 40% for rms wavefront distortion of 0.4λ.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Retrieval of the Wavefront of a Laser Beam Via the Analysis of the in-Focus and Out-of-Focus Intensity Distributions with a Convolutional Neural Network

  • 期刊论文

  • 发表日期：

    2025年4月1日

查看全文
  • A. V. Kotov,
  • Yu. A. Rodimkov,
  • I. B. Meyerov,
  • V. D. Volokitin,
  • S. E. Perevalov,
  • K. F. Burdonov,
  • R. S. Zemskov,
  • A. Soloviev

摘要

We propose a convolutional neural network–based algorithm for retrieving the wavefront shape of a beam using in-focus and out-of-focus intensity distributions. The neural network is trained on numerically generated data. In numerical experiments, the trained model enabled wavefront shape reconstruction with a relative error below 20% for root-mean-square (rms) wavefront distortion in the range of (0.05–0.25)λ, where λ is the radiation wavelength. In a physical experiment conducted on the PEARL laser facility, for a beam with an 18 cm aperture, equal to one-quarter of the focal length, the relative error was approximately 40% for rms wavefront distortion of 0.4λ.