Image reconstruction for wavefront sensing based on ghost imaging and deep learning

A method for Shack Hartmann type wavefront sensor based on ghost imaging has been proposed. In post signal processing for image reconstruction of the proposed methods computational ghost imaging is employed. The proposed method also seems to be suitable for weak power illumination. Therefore, the method is expected to a part of adaptive optics. In this method, image reconstruction has improved to provide more accurate spot patterns with the help of deep learning. From numerical and experimental verifications, the proposed procedure for reconstruction is confirmed to be more effective for improvement of wavefront sensing than the conventional one.