Image deblurring method based on redundant dictionary pair joint optimization

Abstract

The invention discloses an image deblurring method based on a redundant dictionary pair joint optimization, which is used for solving the technical problem that the recovery image in the dual-dictionary sparse image deblurring method under the existing sparse theory framework has strong ring effect at the strong edge. According to the method, a joint optimization function, an iteration solution point spread function and a clear image are established on the basis of the characteristic that the sparse coefficient of a fuzzy image under a fuzzy redundant dictionary is consistent with the sparse coefficient of a clear image under a clear redundant dictionary. The method enhances the robustness against noise, avoids the ill condition in the deconvolution process, reduces the ring effect of the recovery image at the strong edge, and obtains a clearer and more detailed image.

Description

technical field [0001] The invention relates to an image deblurring method, in particular to an image deblurring method based on joint optimization of redundant dictionary pairs. Background technique [0002] The document "Dual Dictionary Sparse Restoration of Motion Blur Degraded Image, Optical Precision Engineering, 2011, Vol19(8), p1982-1989" discloses a double dictionary sparse image deblurring method under the framework of sparse theory. This method first establishes the sparse Transform the degradation and restoration model, then use the redundant dictionary of Haar wavelet coefficients to sparse the image, and finally, use the sparse threshold iterative algorithm to converge the fuzzy image to obtain the restoration image. This method has a good restoration effect on blurred and degraded images. It not only effectively removes motion blur and noise, but also preserves edge details to a certain extent. The method described in the literature is based on the principle o...

AI Technical Summary

Problems solved by technology

The method described in the literature is based on the principle of deconvolution to restore the image. Due to the pathological nature of the deconvolution problem, it is easy to cause the final result to have a ringing effect on the strong edge, which seriously affects the final restoration result.

Images