Spatial domain and transform domain fused noise image reconstruction method

Abstract

The invention provides a spatial domain and transform domain fused noise image reconstruction method. On the basis of respectively improving and optimizing spatial domain bilateral filtering and transform domain wavelet transform shrinkage, the two improved methods are fused, so that image noise can be well removed, detail information and edge features of the image can be reserved, and meanwhile,a ringing phenomenon cannot be generated. Through multi-test argumentation, three iterations are the optimal iteration times of the algorithm; the result obtained by the algorithm provided by the invention achieves a better image denoising effect than that obtained by a BM3D algorithm; compared with the prior art, the method has the advantages that the image edge feature information is reserved, the detail information of the image is enhanced, the identification degree of the image is greatly improved, the noise image is subjected to high-quality denoising reconstruction, and the reconstructedimage has a better visual effect.

Description

technical field [0001] The present invention relates to a noise image reconstruction method, in particular to a noise image reconstruction method combining a space domain and a transform domain, and belongs to the technical field of noise image denoising reconstruction. Background technique [0002] With the rapid development of computer and network information technology, the current society has entered a stage of highly informatized big data, and the forms of information include data, images, videos and other forms. According to statistics, 65% of all kinds of information and data acquired by humans come from images. Digital image technology has achieved great development and extended to various fields, playing a vital role. For example, doctors diagnose diseases of patients through B-ultrasound images, geological exploration teams analyze mineral distribution through satellite remote sensing images, and traffic police monitor and manage urban traffic through surveillance...

AI Technical Summary

Problems solved by technology

[0012] One is that the spatial domain bilateral filtering of the prior art maintains the image edges with strong contrast, but it is easy to introduce noise when maintaining the details of weak contrast in the image; the wavelet transform in the transform domain shrinks and maintains the details of weak contrast images, but it will produce ringing phenomenon. There are obvious defects in the mainstream methods. It is necessary to solve the lack of methods for improving the optimization and fusion of bilateral filtering in the spatial domain and wavelet transform shrinkage in the transform domain. It is impossible to remove image noise while retaining image detail information and edge features. At the same time, there will be no ringing phenomenon

Existing technologies cannot solve these multiple problems at the same time, the method has large limitations, and the application value is not high

[0013] The second is that the number of iterations of the denoising method in the prior art is unscientific. If the number of iterations is too high, the effect of image denoising will be improved very little, but the computational complexity of the algorithm will be increased. If the number of iterations is too small, the effect of image denoising will be too poor. To better solve the problem of denoising accuracy from all aspects, the algorithm time will generally be prolonged, which makes these two problems contradictory. Moreover, there are a large amount of low-frequency noise in the algorithm of the prior art, and the algorithm implementation is very complicated. , it is impossible to enhance the image detail information while retaining the image edge feature information, the recognition degree of the reconstructed noisy image is very low, and the visual effect of the reconstructed image is poor

[0014] The third is that the smoothing effect of the binocular stereo vision matching method in the prior art is poor, and the problem of the critical value of wavelet shrinkage cannot be handled well. The post-image denoising effect is not obvious, and too much noise remains; if the threshold value of wavelet shrinkage is too large, the detailed information of the image will be lost while removing the noise, so that the image obtained after the coefficient reconstruction becomes blurred and the image is lost. Edge detail features will also cause large deviations, resulting in unsatisfactory visual effects of the image

The incorrect selection of the wavelet shrinkage critical value in the prior art has a significant impact on the follow-up work of image denoising

[0015] Fourth, the hard threshold and soft threshold methods of the prior art have obvious shortcomings in image denoising. The overall image of the soft threshold function is blurred, and the image processed by the hard threshold function can better maintain the image details. It is not ideal at the edge. Due to the characteristics of the hard threshold function, the continuity of the image is relatively poor; the threshold function of the prior art cannot be satisfactorily preserved in both the details and the edge, and the recovery effect of the image after denoising is not good. , the complexity of the algorithm is high, it is not easy to implement and the effect is poor, and the robustness of the algorithm is very poor

Images