抑制SAR图像相干斑的迭代方向滤波算法

发布时间：2018-03-01 17:09

本文关键词： 合成孔径雷达图像相干斑迭代方向滤波(IDF) 边缘强度映射(ESM) 各向异性高斯核(AGK)　出处：《电子与信息学报》2014年01期 　论文类型：期刊论文

【摘要】：为保护SAR图像边缘特征并有效提高对乘性相干斑噪声的抑制性能,该文提出一种基于迭代方向滤波的抑制图像相干斑新算法。该算法先借助高斯-伽马平行窗估计出的比率边缘强度映射(ESM)与方向信息,自适应地控制各向异性高斯核(AGK),生成沿ESM方向分布的具有各向异性支撑区域的局域窗。然后将SAR图像多种局部统计参量联合作为衰减因子,形成与SAR图像区域分布特性相适应的负指数衰减型加权系数,进而将负指数衰减型加权系数与局域窗带方向的各向异性支撑区域结合形成局域加权的方向滤波。最后对SAR图像迭代地进行方向滤波即可实现带边缘保护的相干斑抑制。实验结果表明,与多种抑斑算法相比,该文算法在SAR图像抑斑与边缘保护方面均获得了更好的性能。
[Abstract]:To preserve the SAR image edge characteristics and effectively improve the suppression performance of multiplicative speckle noise, this paper proposed a new algorithm to suppress image speckle filter based on iterative direction. The algorithm first coherent with Gauss gamma parallel window ratio edge strength map estimate (ESM) and direction information, adaptive control of anisotropy Gauss (AGK), specific nuclear generating distribution along the ESM direction with anisotropic support domain. Then the local window SAR image obtained by combining several local statistics as a negative exponential attenuation factor, form and adapt to the characteristics of regional distribution of SAR Image Attenuation weighted coefficient, and the negative exponential decay coefficient weighted direction and the local window with the combination of anisotropic directional filtering to form local weighted regional support. At the end of the SAR image iteratively by direction filter can realize speckle suppression with edge protection. The experimental results It shows that the proposed algorithm has better performance in SAR image suppression and edge protection compared with a variety of anti speckle algorithms.

【作者单位】：西安工程大学电子信息学院;西安电子科技大学雷达信号处理国家重点实验室;
【基金】：国家自然科学基金(61271295) 陕西省教育厅项目(12JK0357) 西安工程大学博士科研启动基金(BS1205) 西安市科技局技术转移促进工程项目(CXY1349(2))资助课题
【分类号】：TN957.52

【参考文献】