基于幅度与梯度综合信息的SAR图像非线性扩散去噪方法

探讨SAR图像相干斑抑制的非线性扩散方程方法.以抑制SAR图像噪声,提高图像质量.通过分析SAR图像的幅度分布特性,建立基于SAR图像幅度信息的非线性扩散方程,使方程在幅值较小的背景区域具有较大的光滑作用以抑制噪声,而在幅值较大的目标区域光滑作用较小以保护目标特征.同时为避免噪声对图像中幅度分布的影响,在每一步迭代之前,采用基于梯度信息的非线性扩散方程对图像进行预处理,得到了一种基于图像幅度和梯度综合信息的非线性扩散去噪方法.计算结果表明,本文方法在整体上均具有较好的去噪效果,去噪后的图像较传统方法具有更高的等效视数和边缘保持指数,既能充分抑制背景区域的噪声,又能保护目标点,还很少出现虚假目标.

Nonlinear Diffusion Equation for Image Denoising Based on SAR Amplitude and Gradient Information

This paper discusses the problem of speckle reducing of SAR image based on nonlinear diffusion equation in order to reduce the noise of SAR image and improve the image's quality.It analyses the amplitude distribution of SAR image,and obtain the diffusivity of nonlinear diffusion equation based on this.The new equation has large diffusivity at the background region with small amplitude to reduce speckle efficiently,but has small diffusivity at the target region with large amplitude to protect targets.To avoid the effect of speckle to amplitude distribution,it uses a nonlinear diffusion equation based on gradient to pre-process the speckle in image before each iteration step,thus get a nonlinear diffusion model for speckle reducing based on SAR amplitude and gradient information.Numerical results show that the images processed by the new model have higher ENL(Equivalent Number of Looks) and higher EKI(Edge Kept Index),and it can reduce the speckle at background region efficiently and hold target feature well without the emergence of false targets.