A Fuzzy Kernel Estimation Method Based on l0 and l1 Regular Terms

Abstract

The present invention provides a fuzzy kernel estimation method based on L0 and L1 regular terms. The L0 regular term is used as a constraint when solving the intermediate clear image to effectively suppress the ringing effect. The L1 regular term is used as a regular when estimating the fuzzy kernel. constraint to make the estimated blur kernel sparse enough. According to the established blur kernel estimation model, the blur kernel estimation method based on L0 and L1 regular terms is applied to the image pyramid theory, from the top layer of the image pyramid to the top layer of the image pyramid from coarse to fine. The bottom layer solves and estimates the blur kernel, and uses the obtained blur kernel for the image non-blind restoration algorithm to restore the final clear image. The present invention can more accurately estimate the blur kernel of the blurred image, thereby using the estimated blur kernel of the blurred image to restore the initial blurred image to a final clear image.

Description

technical field [0001] The invention relates to the field of image processing, in particular to a blur kernel estimation method based on L0 and L1 regular terms. Background technique [0002] Ideally, the photos taken by the camera equipment are clear and ideal images, but the images are often affected by various other factors during the imaging process, such as camera shake, the relative distance between the camera equipment and the shooting target, etc. Movement, etc., resulting in the degradation of the quality of the captured image, this process is usually referred to as the image degradation process. The image degradation process will not only lead to a serious decline in the visual effect of the image, but also greatly reduce the practical application value of the image. [0003] Image restoration is to follow the inverse process of image restoration and use the prior knowledge of the image degradation process to obtain a clear image. Although the causes of image deg...

AI Technical Summary

Problems solved by technology

[0006] In order to solve the problem that a single first-order regularization constraint is imposed on the image at present, or the blur kernel estimation is not performed according to the scale of the image, resulting in the inability to accurately estimate the blur kernel, the present invention provides a blurring method based on L0 and L1 regularization terms Kernel estimation methods, including:

Images