An image restoration method under Cauchy noise and its application

Abstract

The invention belongs to the technical field of image processing, and discloses an image restoration method under Cauchy noise and its application. Obtain a degraded image composed of the original image and Cauchy noise; for example, the size of the Cauchy noise determines the size of the scale parameter in the Cauchy distribution; determine the corresponding data fidelity item, and use the mixed regularization scheme to establish the minimization under the Cauchy noise Regular model; under the framework of the alternating direction multiplier method, a fast algorithm is proposed to solve the model, such as the stable solution of the model to obtain the restored image. The invention can effectively remove the blur and Cauchy noise in the image, and the new mixed regularization scheme can overcome the deficiency under the single regularization, reduce or avoid the step effect, maintain the smoothness and detail texture characteristics of the image, and can greatly reduce the time spent on restoring the image. time. After numerical simulation experiments, it has shown advantages in peak signal-to-noise ratio, structural similarity, and running time, and can also see a good restoration effect in terms of visual perception.

Description

technical field [0001] The invention belongs to the technical field of image processing, and in particular relates to an image restoration method under Cauchy noise and its application. Background technique [0002] At present: in the digital age, images are an important aspect for people to understand the objective world, and high-quality images can provide people with more useful information. However, due to the interference and influence of objective conditions such as imaging equipment and the living environment, the obtained images are often degraded and degraded, accompanied by certain noise and blur. In order to further obtain high-quality images, on the one hand, people improve the sensor level of imaging equipment from the hardware level, and on the other hand, they digitize the degraded images from the software level. In recent years, image restoration technology has been widely used and deeply researched in the fields of remote sensing satellite imaging, medical ...

AI Technical Summary

Problems solved by technology

For the parameters in the model, adaptive adjustment can be applied to a wider range of degraded images, but this requires the model to extensively learn the texture structure of natural images, which greatly increases the difficulty and computational complexity of the model solution.

Images