Deep learning wavefront sensing method for Shack-Hartmann sensors with sparse sub-apertures

Direct wavefront sensing with a plenoptic sensor based on deep learning

Traditional plenoptic wavefront sensors (PWS) suffer from the obvious step change of the slope response which leads to the poor performance of phase retrieval. In this paper, a neural network model combining the transformer architecture with the U-Net model is utilized to restore wavefront directly from the plenoptic image of PWS. The simulation results show that the averaged root mean square error (RMSE) of residual wavefront is less than 1/14λ (Marechal criterion), proving the proposed method successfully breaks through the non-linear problem existed in PWS wavefront sensing. In addition, our model performs better than the recently developed deep learning models and traditional modal approach. Furthermore, the robustness of our model to turbulence strength and signal level is also tested, proving the good generalizability of our model. To the best of our knowledge, it is the first time to perform direct wavefront detection with a deep-learning-based method in PWS-based applications and achieve the state-of-the-art performance.

Deep learning assisted plenoptic wavefront sensor for direct wavefront detection

Traditional plenoptic wavefront sensors (PWFS) suffer from the obvious step change of the slope response, leading to poor wavefront detection performance. In order to solve this problem, in this paper, a deep learning model is proposed to restore phase maps directly from slope measurements of PWFS. Numerical simulations are employed to demonstrate our approach, and the statistical residual wavefront root mean square error (RMSE) of our method is 0.0810 ± 0.0258λ, which is much superior to those of modal algorithm (0.2511 ± 0.0587λ) and zonal approach (0.3584 ± 0.0487λ). The internal driving force of PWFS-ResUnet is investigated, and the slope response differences between sub-apertures and directions are considered as a probably key role to help our model to accurately restore the phase map. Additionally, the robustness of our model to turbulence strength and signal-to-noise ratio (SNR) level is also tested. The proposed method provides a new direction to solve the nonlinear problem of traditional PWFS.

Predictive correction method based on deep learning for a phase compensation system with frozen flow turbulence

We propose a deep learning method that includes convolution neural network (CNN) and convolutional long short-term memory (ConvLSTM) models to realize atmospheric turbulence compensation and correction of distorted beams. The trained CNN model can automatically obtain the equivalent turbulent compensation phase screen based on the Gaussian beams affected by turbulence and without turbulence. To solve the time delay problem, we use the ConvLSTM model to predict the atmospheric turbulence evolution and acquire a more accurate compensation phase under the Taylor frozen hypothesis. The experimental results show that the distorted Gaussian and vortex beams are effectively and accurately compensated.

A Method Used to Improve the Dynamic Range of Shack-Hartmann Wavefront Sensor in Presence of Large Aberration

With the successful application of the Shack-Hartmann wavefront sensor in measuring aberrations of the human eye, researchers found that, when the aberration is large, the local wavefront distortion is large, and it causes the spot corresponding to the sub-aperture of the microlens to shift out of the corresponding range of the sub-aperture. However, the traditional wavefront reconstruction algorithm searches for the spot within the corresponding range of the sub-aperture of the microlens and reconstructs the wavefront according to the calculated centroid, which leads to wavefront reconstruction errors. To solve the problem of the small dynamic range of the Shack-Hartmann wavefront sensor, this paper proposes a wavefront reconstruction algorithm based on the autocorrelation method and a neural network. The autocorrelation centroid extraction method was used to calculate the centroid in the entire spot map in order to obtain a centroid map and to reconstruct the wavefront by matching the centroid with the microlens array through the neural network. This method breaks the limitation of the sub-aperture of the microlens. The experimental results show that the algorithm improves the dynamic range of the first 15 terms of the Zernike aberration reconstruction to varying degrees, ranging from 62.86% to 183.87%.

Phase-diversity wavefront sensing enhanced by a Fourier-based neural network

Phase diversity wavefront sensing (PDWS) has been a successful approach to quantifying wavefront aberrations with only a few intensity measurements and nonlinear optimization. However, the inherent non-convexity of the inverse problem may lead to stagnation at a local minimum far from the true solution. Proper initialization of the nonlinear optimization is important to avoid local minima and improve wavefront retrieval accuracy. In this paper, we propose an effective neural network based on low-frequency coefficients in the Fourier domain to determine a better estimate of the unknown aberrations. By virtue of the proposed network, only a small amount of simulation data suffice for a robust training, two orders of magnitude less than those in existing work. Experimental results show that, when compared with some existing methods, our method achieves the highest accuracy while drastically reducing the training time to 1.4 min. The minimum, maximum, and mean values of the root mean square (RMS) residual errors for 800 aberrations are 0.017λ, 0.056λ, and 0.039λ, respectively, and 95% of the RMS residual errors are less than 0.05λ.

Speckle denoising based on deep learning via a conditional generative adversarial network in digital holographic interferometry

Speckle denoising can improve digital holographic interferometry phase measurements but may affect experimental accuracy. A deep-learning-based speckle denoising algorithm is developed using a conditional generative adversarial network. Two subnetworks, namely discriminator and generator networks, which refer to the U-Net and DenseNet layer structures are used to supervise network learning quality and denoising. Datasets obtained from speckle simulations are shown to provide improved noise feature extraction. The loss function is designed by considering the peak signal-to-noise ratio parameters to improve efficiency and accuracy. The proposed method thus shows better performance than other denoising algorithms for processing experimental strain data from digital holography.