Far-field phasing method based on deep learning for tiled-aperture coherent beam combination

In this work, a phase measurement method based on deep learning is proposed, which uses the deep convolutional neural network algorithm to measure the phase error of tiled-aperture coherent beam combination system in the far field. Fig. 1.tiff describes the simulated two-beam far-field interference pattern and the corresponding axial intensity distribution with different phase errors. Fig.2.tiff describes the comparison of simulation results. Fig.3.tiff describes the four-beam far-field interference pattern and the application of amplitude modulation. Fig.4.tiff describes the results of four beam simulation. Fig.5.tiff describes the experimental setup and DCNN training process. Fig.6.tiff describes experimental data and experimental results.