A Nonlocal Mean Filtering Method for Synthetic Aperture Radar Images

Abstract

The invention provides a synthetic aperture radar image non-local mean filtering method, which belongs to the field of radar image processing. The invention introduces heterogeneity measurement into the non-local mean algorithm, and improves the weight measurement method of the non-local mean algorithm. First of all, in view of the traditional non-local mean method using equal weight weighting method in calculating the distance measure, this paper introduces the weighted weight value of the distance measure of the pixel point in the image block window by introducing the coefficient of variation, because the coefficient of variation can represent the texture direction of the image , so the introduction of the image block similarity weighted by the coefficient of variation can more effectively capture the image blocks with similarity in texture and direction, thereby avoiding the blurring of edges and directions; then, a method based on the coefficient of variation is designed The self-adaptive degradation parameter function, adaptive adjustment of the degradation parameters, can effectively ensure that the heterogeneous area is protected from being over-smoothed. Therefore, the present invention can more effectively maintain texture edge information while filtering coherent speckles.

Description

technical field [0001] The invention belongs to the field of radar image processing, in particular to a non-local mean value filtering method of a synthetic aperture radar image. Background technique [0002] Synthetic Aperture Radar (SAR) is an active high-resolution imaging sensor that can observe all-weather and all-weather, and has strong penetration. It is widely used in military reconnaissance, disaster monitoring, and surface cover detection. and other fields. However, because SAR imaging uses echo coherent superposition for imaging, coherent speckle noise inevitably exists in SAR images. These coherent speckle noises reduce the visual effect of SAR images and bring difficulties to subsequent image segmentation and target recognition. Therefore, the speckle suppression of SAR images is of great significance to the segmentation, classification and recognition of SAR images. [0003] Speckle filtering methods for SAR images can be divided into spatial domain filtering...

AI Technical Summary

Problems solved by technology

However, when these methods calculate the distance measure, they use the mean value of the gray distance of the pixels in the window, or the Gaussian weighted value, without considering the texture direction information of the image, and there is a large amount of texture direction information in the SAR image. This information should be an important basis for distance measurement

In addition, in the NLM algorithm, the weight is a Gaussian kernel function that depends on the distance measure, and the degradation parameter has a very important influence on the filtering effect. However, there are few literatures on the degradation parameter, which are generally set according to the empirical value. Does not achieve the effect of adaptive denoising

Images