.png){kind=icon}
.png){kind=icon}
Return
Robust wavefront sensing method based on machine learning for optical thickness profiling
.png){kind=icon}
PRE
AI
DOI：10.1007/s12206-025-2508-z.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
In this research, a new machine learning-based wavefront sensing method is proposed to measure the optical thickness variations of the transparent parallel glass using only single fringe pattern. Data generation was conducted using the Zernike polynomials and corresponding coefficients, and the training network was developed by combining a U-shaped structure with robust convolution neural network (CNN) architecture. The performance evaluation of the proposed method was performed with simulation data, and experimental verification of the proposed method also conducted by comparing the measurement results of the proposed method for optical thickness variations and those of existing phase-extraction methods. The proposed method using a single fringe pattern does not incur systematic errors caused by phase-shifting nonlinearity and has measurement uncertainty as 0.89 nm which is significantly lower than that of the conventional methods.
Keywords:
Fringe pattern
Interferometry
Machine learning
Optical thickness
Wavefront sensing
Journal
IF:
1.7
Papers:
601
Citations:
1.2W