Phase unwrapping is a classical signal processing problem, which refers to the recovery of the original phase value from the wrapped phase. Two dimensional phase unwrapping is widely used in optical measurement technology, such as digital holographic interferometry, fringe projection profilometry, synthetic aperture radar and many other applications. In this paper, a phase unwrapping method with the convolution neural network is proposed, and the feasibility is analyzed by numerical simulation. The convolution neural networks with different parameters are set up, and the phase screens used for the training set and testing set of convolution neural network are simulated with MATLAB software. The numerical simulation results show that the four convolution neural network models can be used for phase unwrapping, but the parameters have a significant impact on its accuracy.
For adaptive optics without wavefront detection, the wavefront control method based on deep learning is analyzed. The simulation model of adaptive optics is established,The far-field spot data collected by the photodetector is used as the input of the neural network model, and the Zernike mode coefficient is used as the output. The fully trained model can quickly and accurately recover and control the low-order wavefront. The simulation results show that convolution neural network can effectively extract image features, which is better than ordinary depth neural network model. For convolution network model, the larger the number of training sets, the smaller the value of loss function after convergence, and the higher the accuracy of the model. Compared with the traditional iterative optimization control method, the control method based on neural network model has obvious advantages in real-time.
The phase grating wavefront curvature sensor based on liquid crystal spatial light modulator is introduced. A close-loop phase retrieval method based on Eigen functions of Laplacian is proposed, and its accuracy and efficiency are analyzed through numerical experiments of atmospheric phase retrieval. The results show that the close-loop phase retrieval method has a high accuracy. Moreover, it is stable regardless of modal cross coupling.
Considering the influences of speckle noises and wavefront aberration error on image quality in active imaging based on spatial heterodyne detection, a wavefront correction method that based on metric optimization of multi images is proposed, in which the multi images are generated by aperture dividing technique. An experimental setup is established, and the experiments that correcting the aberration of itself with the above method are performed, in which the stochastic parallel gradient descent algorithm and image sharpness function are used. The results show that the method of multi images averaging can be used to improve the signal-to-noise ratio of target image effectively, and a higher quality of target image can be achieved after correction by optimizing the image sharpness metric that generated with the averaging data of multi images.