摘要
提出一种基于最大后验和非局域约束的非下采样轮廓波变换域SAR图像去噪方法。根据SAR图像数据的特征,引入非对数加性模型,并在该模型下对SAR图像NSCT域中的噪声分布统计建模,应用最大后验准则和非局域约束相结合的方法解求SAR图像真实信号的NSCT系数。试验结果表明,本方法具有良好的去噪能力。
An algorithm of SAR image denoising in nonsubsampled contourlet transform(NSCT) domain based on maximum a posteriori(MAP) and non-local(NL) restriction is proposed.Firstly a nonlogarithmic additive model is applied to the SAR image data,and then the statistical distribution of the noise within the nonlogarithmic additive model in the SAR image is modeled in the NSCT domain.MAP and NL restriction are used to get the NSCT coefficients of the real signal.Experiments show that the algorithm proposed is effective in SAR image denoising,and also possess high performance over many traditional algorithms.
出处
《测绘学报》
EI
CSCD
北大核心
2012年第1期59-64,共6页
Acta Geodaetica et Cartographica Sinica
基金
国家863计划(2007AA120203)
国家973计划(2011CB707103)
国家自然科学基金(40930532)
关键词
SAR图像去噪
非对数加性模型
非下采样轮廓波
最大后验
非局域
SAR image denoising
nonlogarithmic additive model
nonsubsampled contourlet transform(NSCT)
maximum a posteriori(MAP)
non-local(NL)