Fractional order adaptive coherent speckle filtering method based on image form fuzzy membership degree

Abstract

The invention discloses a fractional order adaptive coherent speckle filtering method based on an image form fuzzy membership degree. According to the method, alternate iteration of two steps of fractional order total variation regularization additive noise reduction and residual image weighted feedback is performed to realize coherent speckle filtering; estimations of noise standard deviation, cartoon image components and local variance of a corresponding residual image are utilized to calculate fuzzy membership degrees of each pixel point in three forms of image border, grain and smoothness; on this basis, an adaptive calculation method of model parameters is provided; and the calculation of fractional order difference is simplified, and the fractional order adaptive coherent speckle filtering method is provided. By using the method, noise and a staircase effect can be effectively suppressed, the image border and grain details can be better kept, and a filtered image has a good visual effect. The method has the advantages of high calculation speed and good practicability in adaptive calculation of arithmetical parameters, and has wide application prospects in the fields of remote sensing, synthetic aperture radars, medical imaging and the like.

Description

Technical field [0001] The present invention belongs to the field of image processing, in particular to a filtering technique for suppressing image multiplicative noise (coherent speckle), in particular to a fractional-order adaptive coherent speckle filtering method based on the fuzzy membership of image morphology. Background technique [0002] In the coherent imaging process such as remote sensing, synthetic aperture radar, and nuclear magnetic resonance, the image is inevitably contaminated by multiplicative noise (coherent speckle). Coherent speckle noise seriously reduces the image quality and affects the interpretation, classification and further processing of the image. Therefore, the suppression of coherent speckle noise is a necessary step before many image post-processing. In the process of suppressing image coherent speckle noise, it is very important to maintain important geometric structures such as image edges and textures. At the same time, in order to adapt to t...

AI Technical Summary

Problems solved by technology

However, this method can only solve the problem of additive noise denoising, and has not been used for coherent speckle multiplicative noise suppression.

[0006] In addition, although in Zhang Jun’s two articles based on the fractional total variation image denoising method, the model parameters have been adaptively calculated, and the image texture, cartoon (including smooth and edge ) morphological components, but the parameter estimation method used in this article is affected by the texture, which is prone to large errors. For the classification of image texture morphology and cartoon morphological components, as well as the calculation of fractional difference orders, simple hard methods are used. The threshold method can easily lead to inaccurate classification, and the gray level of the processed image has obvious steps at the junction of different image morphological components, and the visual effect is not ideal.

Images