摘要
二元收缩方程定义了由相邻尺度小波系数的联合概率密度函数,其与噪声模型联立后利用最大后验概率估计可进行图像去噪。在SAR图像斑点噪声服从瑞利分布的假设下,结合双树复数小波变换推导了基于二元收缩方程的SAR图像的简化去噪模型,然后利用局部方差估计和维纳滤波器获得噪声方差与带噪小波系数方差的估计值,并计算出合适阈值对SAR图像进行去噪。实验结果显示,去噪图像的峰值信噪比以及有效视数都较其它算法有大幅提高,且很好地保持了图像的边缘特征。
Bivariate shrinkage functions （BSF） statistically denoted as joint probability density functions （PDF） and noise PDF could be united by MAP to de-noise image. The intensity of speckle was hypothesized to be distributed according to Rayleigh distribution. Then SAR image de-noising modal based on BSF and dual-tree complex wavelet transform （DT-CWT） was constructed and reduced. Local variance estimation and wiener filter were used to estimate noise variance and noisy wavelet coefficients variance, which were used to choose an appreciated threshold to de-noise SAR image. Experiment results demonstrate that PSNR and ENL values of de-noised images are extremely larger than other algorithms and edge features have been perfectly preserved.
出处
《系统仿真学报》
CAS
CSCD
北大核心
2008年第1期40-45,共6页
Journal of System Simulation
关键词
二元收缩方程
双树复数小波变换
局部方差估计
SAR图像去噪
bivariate shrinkage functions
dual-tree complex wavelet transform
local variance estimation
SAR image de-noising
