摘要
提出了基于小波域高斯混合模型贝叶斯估计模糊萎缩的SAR图像降斑算法。该算法分析了SAR图像在平稳小波变换(SWT)域中的统计模型,并用高斯混合模型对其进行描述,推导出基于贝叶斯估计的信号最小均方误差(MMSE)的模糊萎缩因子。籍此再根据小波域相邻尺度间小波系数的相关性,采用分区域模糊萎缩思想,很好地得到无斑点真实信号小波系数的估计。仿真结果表明该算法在大大抑制斑点噪声的同时,有效的保持了边缘,其性能优于改进Lee滤波,小波软阈值和SWT萎缩降斑算法。
An efficient despeckling method was proposed based on stationary wavelet translation （SWT） for synthetic aperture radar （SAR） images. The statistical model of wavelet coefficients was analyzed and its performance was modeled with a mixture density of two zero-mean Gaussian distributions. A fuzzy shrinkage factor was derived by employing the minimum mean error （MMSE） criteria with bayesian estimation. Furthermore, the ideas of region division and fuzzy shrinkage were adopted according to the interscale dependencies of the wavelet coefficients. The noise-free wavelet coefficients were estimated accurately. Experimental results show that the method is superior to the refined Lee filter,wavelet soft thresholding shrinkage and SWT shrinkage algorithms in terms of smoothing effects and edges preservation.
出处
《电波科学学报》
EI
CSCD
北大核心
2006年第6期944-949,共6页
Chinese Journal of Radio Science
基金
中国博士后科学基金(J63104020156)
国防重点实验室基金(51431020204DZ0101)
