Weak light speckle imaging recovery method based on deep convolutional generative adversarial network

Abstract

The invention provides a weak light speckle imaging recovery method based on a deep convolutional generative adversarial network. The method comprises the following steps: S1, obtaining a speckle PSF of a point light source; S2, obtaining speckles I of an unknown object; S3, performing image gray-scale adaptive nonlinear normalization on the speckle I of the unknown object and the speckle PSF of the point light source to obtain an image; S4, performing deconvolution operation on the normalized speckle of the unknown object according to the most approximate noise-to-signal ratio of a scatterer imaging system and the normalized speckle of the point light source to obtain a restored image Otem of the unknown object; and S5, inputting the unknown object recovery image Otem into a pre-trained deep convolutional generative adversarial network model to obtain a final unknown object reconstruction image O. A complete closed-loop speckle recovery imaging method can be constructed from information optics, adaptive optimization and deep learning, the ability of understanding convolution speckle imaging is enhanced, and generalization of deep learning in speckle imaging recovery is greatly improved.

Description

technical field [0001] The invention relates to the technical field of computational optical imaging, and more specifically, to a method for restoring weak-light speckle imaging based on a deep convolutional generative adversarial network. Background technique [0002] Optical imaging is one of the important ways for human beings to receive information. In scenarios such as astronomical telescopes, security cameras, biomedical imaging, and microscopic imaging, due to the existence of scattering media such as smog, rain, snow, and biological tissues, light is scattered by particles of the wavelength order of magnitude in the medium, resulting in incident light The initial wavefront is severely distorted, so it is difficult to obtain high-quality images only by conventional imaging methods. Although the incident light carrying the original image is scattered into seemingly irregular speckles, the information it carries is not lost, but modulated by the scattering medium. The...

AI Technical Summary

Problems solved by technology

[0006] The present invention provides a low-light speckle imaging recovery method based on deep convolution generation confrontation network, which solves the disadvantage of low speckle-related imaging quality under low-light conditions in the prior art

Images