PSF (point spread function) estimation method for parameterized simple lens

Abstract

The invention discloses a PSF (point spread function) estimation method for a parameterized simple lens, and belongs to the field of image processing. The method comprises the following steps: capturing a fuzzy image through a simple lens, converting the PSF estimation of the simple lens into a blind convolution image restoration problem, constructing a target function, constructing a two-dimensional Gaussian distribution function for simulating the PSF of the simple lens, taking the two-dimensional Gaussian distribution function as an initial value of the blind convolution image restoration target, solving a target function through an optimization algorithm, and obtaining a final PSF. According to the invention, the constructed Gaussian distribution function can simulate the PSF distribution conditions of spherical aberration and astigmatism in a simple lens system in a better way through the combination of different parameters, enables a PSF iterative optimization initial value to approach to a real PSF value as much as possible, thereby shortening the iterative optimization time, avoiding the partial optimal solution, and improving the PSF estimation precision.

Description

technical field [0001] The invention relates to image processing technology, in particular to a parameterized simple lens PSF estimation method. Background technique [0002] In recent years, computational imaging with simple lenses has gradually become one of the research directions in the field of image processing. Simple lens front-end imaging equipment usually includes 1-3 lenses, combined with post-image restoration algorithms to obtain high-quality images. Simple lens computing imaging equipment is simple, small in size and low in cost, and plays an important role in high-altitude satellite imaging and large field of view imaging. [0003] The key of simple lens image restoration is to accurately estimate the PSF of the simple lens first, and then use a certain image restoration algorithm to obtain a clear image based on the obtained PSF. The existing simple lens PSF estimation methods usually select a 3×3 Gaussian function or a delta function as the initial value of...

AI Technical Summary

Problems solved by technology

[0004] In order to overcome the above-mentioned disadvantages, the present invention aims to provide a parameterized simple lens PSF estimation method, and constructs a parameterized two-dimensional Gaussian function to simulate the initial PSF value of a simple lens, so as to solve the problem that the PSF initial value and the real PSF value in the prior art are too different. Large, resulting in long iteration time, local solutions, and low PSF estimation accuracy.

Images