An Image Deblurring Method Based on Fourier Transform

Abstract

The invention discloses an image defuzzification method based on Fourier transform. The method comprises the following steps: 1, by use of a k-svd algorithm, pre-denoising an input fuzzy image; 2, initializing a fuzzy kernel operator and dimension parameters; and 3, by taking change values of the dimension parameters as iteration termination conditions, performing internal iteration operation on the pre-denoised fuzzy image so as to obtain an image matrix after processing, according to the processed image matrix, updating the fuzzy kernel operator, and by taking preset frequency as an iteration termination condition, by use of the updated fuzzy kernel operator, performing external iteration operation on the processed image matrix so as to obtain a final denoised image. According to the image defuzzification method, the image is denoised first of all, then an estimation result is continuously optimized by use of an iteration method, and rapid convolution operation is finished by use of the Fourier transform, such that the calculation efficiency is improved, and too much prior arts about the image are unnecessary.

Description

technical field [0001] The invention relates to the field of computer image processing, in particular to an image deblurring method based on Fourier transform. Background technique [0002] Image blurring caused by factors such as motion and camera shake is a common phenomenon nowadays, and the blurring directly leads to the degradation of image quality. In addition, under low-light conditions, camera imaging often requires a longer exposure time, which leads to slight hand shakes that will seriously affect the final image quality. Therefore, in real life, image deblurring is necessary and has practical application value. [0003] In recent years, many deblurring algorithms have been proposed. While estimating the original image, the corresponding blur kernel operator is also estimated. Xue-fen, Yi et al. used a point spread function to estimate, mainly using the function to deconvolute the blurred image; Jiun-Lin, Chia-Feng et al. proposed a statistical method based on th...

AI Technical Summary

Problems solved by technology

Xue-fen, Yi et al. used a point spread function to estimate, mainly using the function to deconvolute the blurred image; Jiun-Lin, Chia-Feng et al. proposed a statistical method based on the predicted image and the blurred image The iterative method of features achieves the effect of deblurring, but the disadvantage is that there will be "ringing phenomenon" and it takes a long time; Zohair, Ghazali et al. proposed a deblurring method based on Laplacian filtering, which is convenient and simple to calculate , but the effect is not significantly improved

Images