Super-lens chromatic aberration recovery method based on deep learning

Abstract

The invention discloses a super-lens chromatic aberration recovery method based on deep learning. The method comprises the following steps: 1) constructing a super-lens chromatic aberration recovery neural network and a chromatic aberration recovery loss function; 2) taking the super-lens chromatic aberration image x and a standard image y corresponding to the super-lens chromatic aberration image x as a training data pair, training the super-lens chromatic aberration recovery neural network to enable a chromatic aberration recovery loss function value to be less than 8, realizing bidirectional mapping from the super-lens chromatic aberration image to the standard image corresponding to the super-lens chromatic aberration image, and obtaining a trained super-lens chromatic aberration recovery model; and 3) inputting the to-be-recovered super-lens chromatic aberration image into the trained super-lens chromatic aberration recovery model by using the step 2) to realize super-lens chromatic aberration recovery.

Description

technical field [0001] The invention relates to the field of digital image processing, in particular to a method for restoring chromatic aberration of a super-lens based on deep learning. Background technique [0002] A metalens is a flat lens based on a metasurface. With its outstanding features such as ultrathinness and low cost, it has become a breakthrough technology in the development of micro-optical systems. However, due to the high resonance phase dispersion of each nano-element in the metalens structure and the inherent dispersion of the materials used, the metalens has been suffering from huge chromatic aberration, which seriously reduces the focusing and imaging quality. In 2015, Zhongyang Li et al published "Visible-frequency metasurfaces for broadband anomalous reflection and high-efficiency spectrum splitting" in "Nano Letters" in 2015, and proposed a broadband anomaly with high conversion efficiency at visible and near-infrared frequencies Reflective metalens...

AI Technical Summary

Problems solved by technology

However, due to the high resonance phase dispersion of each nano-element in the metalens structure and the inherent dispersion of the materials used, the metalens has been suffering from huge chromatic aberration, which seriously reduces the focusing and imaging quality.

In 2015, Zhongyang Li et al published "Visible-frequency metasurfaces for broadband anomalous reflection and high-efficiency spectrum splitting" in "Nano Letters" in 2015, and proposed a broadband anomaly with high conversion efficiency at visible and near-infrared frequencies reflective metalens, but this metalens still cannot solve the problem of chromatic aberration due to the structure

Affected by chromatic aberration, there are problems such as blurred edges and poor image quality in metalens images

In the era of big data, for massive metalens image data, there is a lack of methods that can restore chromatic aberration of images captured by metalens quickly and at low cost.

Images