2022 · SPIE Photonics Europe 2022
Noise robust focal distance detection in laser material processing using CNNs and Gaussian processes
Sepehr Elahi , Can Polat , Omid Safarzadeh , Parviz ElahiResearch highlights
Focus detection with confidence under optical noise
Estimate laser focal distance from noisy diffraction images while reporting confidence in the prediction.
Why we investigated it
Surface roughness and camera noise can change the images used for focus control, so accuracy on clean images is insufficient.
What the work contributes
Fourier-optics simulations separate source-wave noise from detection noise. A CNN and Gaussian-process classifier map the resulting images to focal distances and confidence estimates.
What the evidence shows
The study evaluates predictions across simulated focal distances and noise levels, providing a controlled way to examine noise sensitivity.
Build on this work
Use the simulation and classifier design in the paper to plan noise-controlled focus-detection experiments.
- Laser focus detection
- Gaussian processes
- Uncertainty estimation
- Fourier optics
Abstract
In this work, we investigate the effects of noise on real-time focal distance control for laser material processing by generating the images of a sample at different focal lengths using Fourier optics and then designing, training, and testing a deep learning model in order to detect the focal distances from the simulated images with varying standard deviations of added noise. We simulate both input noise, such as noise due to surface roughness, and output noise, such as detection camera noise, by adding zero-mean Gaussian noise to the source wave and the simulated image, respectively, for different focal distances. We then train a convolutional neural network combined with a Gaussian process classifier to predict focus distances of noisy images together with confidence ratings for the predictions.
Cite
@article{elahi2022noise,
title = {Noise robust focal distance detection in laser material processing using CNNs and Gaussian processes},
author = {Sepehr Elahi and Can Polat and Omid Safarzadeh and Parviz Elahi},
year = 2022,
journal = {SPIE Photonics Europe 2022},
doi = {10.1117/12.2624337},
}