Method for restoring weighting sparse edge regularization image

Abstract

The invention discloses a method for restoring a weighting sparse edge regularization image, which comprises the following steps: (1) using a difference operator to obtain gradient information of each direction of a blurred noise image and combining directions at random so as to obtain a matrix mathematical model of the edge of the image; (2) according to an image sparse principle and the definition of the optimal sparse solution of an uncertain equation, aiming at an image edge model obtained in the step (1), adopting a weighting sparse constraint to constrain the edge of the image, and deriving the cost function of image restoration by combining a known degenerate function; and (3) according to the steps (1) and (2), using an improved upper bound minimal method to carry out optimization solving on the cost function so as to obtain a clear restored image. The method greatly enriches information content at the edge of the image; the weighting sparse edge constraint is adopted for effectively protecting the edge characteristics of the image; and the restored image with high quality is obtained by using the improved upper-bound minimal mathematical method.

Description

technical field [0001] The invention belongs to computer image processing technology and relates to an image restoration method. Background technique [0002] Image degradation is ubiquitous, and it exists in almost all fields related to image processing. In the process of acquiring images, the image blur caused by the mismatch between the object distance and the image distance between the camera device and the object to be photographed is called defocus blur; due to the relative motion between the imaging device and the object to be photographed Image blur is called motion blur. At the same time, the image will also be blurred due to external weather, sunlight, air, dust on the lens of the imaging device, and the hardware of the imaging device itself. In order to mine more and more accurate information from damaged images, a series of image processing technologies such as image clarity restoration and ultra-high resolution reconstruction have been widely studied and succes...

AI Technical Summary

Problems solved by technology

[0005] In the existing regularization methods based on image edge information, most methods only use the horizontal and vertical directions of the image, and the edge information in the two directions is used as the regularization item of the cost function, which not only fails to fully utilize The edge information of the image, on the other hand, ignores the edge characteristics of the image, resulting in insufficient sharpening of the edge of the restored image and low restoration quality.

Images