Phase recovery improvement method and device for autocorrelation signal with interference

Abstract

The invention discloses a phase recovery improvement method and device for an autocorrelation signal with interference, and the method comprises the steps: extracting the initial Fourier amplitude ofa target object from a known target object autocorrelation signal, and obtaining a phase recovery reconstruction result through employing a random Fourier phase as an input; performing spatial domaindenoising processing on the reconstruction result, assigning zero to a part smaller than the maximum energy setting percentage of the reconstruction result to obtain a denoised reconstruction result,and calculating a mean square error between the Fourier amplitude of the denoised reconstruction result and the initial Fourier amplitude of the target object; repeating the above steps by taking a plurality of different random Fourier phases as inputs to obtain a plurality of denoised reconstruction results and a plurality of mean square errors; and taking the mean square error as an evaluation index of the quality of the reconstruction result, selecting an optimal reconstruction result from the plurality of denoised reconstruction results as final output, and realizing phase recovery of thetarget object. According to the invention, stable phase recovery can be realized for an autocorrelation signal with interference.

Description

technical field [0001] The invention relates to the fields of computer vision and digital image processing, in particular to an improved phase recovery method and device for autocorrelation signals with interference. Background technique [0002] The traditional phase recovery algorithm can be used to restore and reconstruct the target object to be observed from the autocorrelation signal of the target object, that is, firstly use the Fourier transform to extract the Fourier amplitude of the target object from the autocorrelation signal; When the leaf amplitude is known and the phase is unknown, the random phase is used as the initial guess of the Fourier phase of the target, and the real Fourier phase of the target is iteratively restored and the spatial distribution of the target is reconstructed. However, due to the randomness of the initial phase guess, the spatial distribution of the target object reconstructed by the traditional phase recovery algorithm and the real Fo...

AI Technical Summary

Problems solved by technology

However, due to the randomness of the initial phase guess, the spatial distribution of the target object reconstructed by the traditional phase recovery algorithm and the real Fourier phase of the target object recovered are very unstable, especially when there is some interference in the autocorrelation signal of the target object When , this reconstruction instability will be greater, leading to confusion in the phase recovery results, and it is difficult to obtain quasi-effective target object space information from a single phase recovery result, which limits the traditional phase recovery algorithm for interference with autocorrelation signals. applicability

Images