Perfect information non-local constraint total variation method for image recovery

Abstract

The invention discloses a perfect information non-local constraint total variation method for image recovery. By the method, the problems that edges cannot be sharpened and high-frequency details cannot be recovered during the image recovery in the prior art are solved. The technical scheme is as follows: (1) suppressing noise of an indefinite image by using a non-local mean value filter method; (2) initializing a recovered result by using a wiener filter method; (3) calculating a perfect information non-local weight coefficient matrix; (4) updating the recovered result by using a threshold value iterative formula; (5) suppressing noise of the recovered result by using a total variation denoising method; (6) judging whether to update the perfect information non-local weight coefficient matrix, if so, returning to the step (3), otherwise, executing step (7); and (7) judging whether a stop condition is met, if so, obtaining a final result, otherwise, returning to the step (4) until the stop condition is met. During recovery, the edge of an image can be sharpened and the high-frequency details of the image can be recovered; and the method can be used for recovering the indefinite image in a known indefinite type.

Description

technical field [0001] The invention belongs to the technical field of image processing, in particular to a method for restoring blurred images, which can be used for restoring blurred images of various known blur types. Background technique [0002] Image restoration refers to the removal or mitigation of image quality degradation in the process of acquiring digital images. It is an important and challenging research content in image processing. For the image restoration problem, researchers have proposed many methods. [0003] Traditional restoration methods include inverse filtering, Wiener filtering, Kalman filtering and generalized inverse singular value decomposition, etc. These methods have been widely used in image restoration, but these methods require blurred images to have a high signal-to-noise ratio , methods such as inverse filtering are only suitable for images with high SNR, which limits the practical application of traditional restoration methods. Another ...

AI Technical Summary

Problems solved by technology

The convergence speed of this method is higher than that of the general threshold iteration method. At the same time, in their code example, J.Bioucas-Dias et al. converted the noise coefficient into the fully variable domain for suppression, and removed the ringing effect. However, this method tends to produce a staircase effect in the smooth area of ​​the image, and cannot restore the high-frequency details of the image very well.

Images