Optimized Approach for Image Design Processing in Optical Networks: A Comparative Study

Abstract

Deep space detection and remote sensing both require optical imaging devices. The optical imaging system often needs a bigger aperture mirror to attain high spatial resolution. As a result, several novel optical imaging systems, such as big segmented mirror telescopes, large aperture membrane diffractive optical telescopes, and others, have been researched in recent years. Real-time wavefront measurement is not required for the wavefront sensorless (WFSless) applied optics (AO) approach. The wavefront corrector is directly regulated via feedback following an image quality measure of the far-field image to correct for wavefront aberration. Integrating artificial neural networks (ANN) and deep learning plays a vital role in developing WFSless AO systems. This paper evaluated various important aspects to provide an in-depth review of the state-of-the-art machine learning-based algorithms deployed in WFSless AO systems. Finally, the applications and prospects were outlined.