Full-Color, Wide Field-of-View Metalens Imaging via Deep Learning

Abstract

Chromatic aberration has been the main showstopper for metalenses when it comes to imaging applications with broadband sources such as ambient light. In wide field-of-view metalenses, this challenge becomes far more severe due to exacerbated lateral chromatic aberrations. In this paper, it is demonstrated, for the first time, full-color wide field-of-view imaging using a fisheye metalens coupled with deep learning computational processing. This approach is capable of restoring panoramic images with enhanced signal-to-noise ratio while effectively correcting chromatic aberration, distortion, and vignetting. Furthermore, it is shown that the deep learning algorithm is robust against various lighting conditions and object distances, making it a versatile solution for practical imaging applications involving wide field-of-view metalenses.