Using the angular spectrum method and Fresnel transform as kernels in a modified Gerchberg-Saxton algorithm

Abstract

This work uses the angular spectrum method and the Fresnel transform as the kernel of a modified Gerchberg-Saxton algorithm (MGSA) for image reconstruction in lensless optical systems using the retrieved phases. The MGSA yielded image reconstructions with high similarity when compared to the known diffraction plane image and assessed using Linfoot's criteria of merit. The relatively low image sharpness is attributed to the sampling issues in the numerical simulation of the angular spectrum method and the Fresnel transform.