An Image Deblurring Method Based on Fourier Transform

Abstract

The invention discloses an image deblurring method based on Fourier transform, which includes: (1) using k-svd algorithm to pre-denoise the input blurred image; (2) initializing a blur kernel operator and scale parameters; (3) Using the change value of the scale parameter as the iteration termination condition, perform an internal iteration operation on the pre-denoised blurred image to obtain the processed image matrix; update the fuzzy kernel operator based on the processed image matrix, and use a predetermined number of times as the iteration termination condition , using the updated blur kernel operator, perform an external iteration operation on the processed image matrix to obtain the final deblurred image. The image deblurring method of the present invention first denoises the image, then uses an iterative method to continuously optimize the estimation results, and uses Fourier transform to complete fast convolution operations, which improves calculation efficiency and does not require too much Prior knowledge about the image.

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