基于NSCT和纹理特征的SAR图像相干斑抑制  

SAR(Synthetic Aperture RADAR) Image Speckle Noise Suppression Based on Non-Subsampled Contourlet Transform(NSCT) and Texture Features

作　　者：沈荻帆 丁洁[1] SHEN Difan;DING Jie

机构地区：[1]海南大学信息与通信学院

出　　处：《科技创新与应用》2019年第33期1-4,8,共5页Technology Innovation and Application

基　　金：海南省自然科学基金面上项目(编号：618MS029)

摘　　要：提出了一种基于非下采样轮廓波变换和纹理特征的合成孔径雷达(SAR)图像相干斑噪声抑制方法。由于SAR图像的相干成像原理,造成其固有的相干斑噪声,对后续的图像处理会造成很大的困难。针对相干斑噪声的抑制问题,利用非下采样轮廓波变换的平移不变特性,结合变换后的高频子带系数的纹理模型特征,生成各子带系数自适应阈值,再经反变换,达到相干斑噪声抑制目的。实验数据表明,该方法经边缘保持指数(ESI)和等效视数(ENL)评测,能较好的抑制噪声,且更好的保持边缘。A method for speckle noise suppression in Synthetic Aperture Radar(SAR)images based on non-subsampled contourlet transform and texture features was proposed.Due to the coherence imaging principle of SAR image,the inherent speckle noise will cause great difficulty to the subsequent image processing.Aiming at the problem of speckle noise suppression,an adaptive threshold of various subband coefficients is generated by combining the texture model features of the transformed high-frequency subband coefficients with the translation invariant characteristics of the non-subsampled contourlet transform.Experimental data show that the proposed method can better suppress noise and maintain edge by Edge Saved Index(ESI)and Equivalent Number of Looks(ENL).

关 键 词：合成孔径雷达 非下采样轮廓波变换 等效视数 边缘保持指数 相干斑噪声抑制 

分 类 号：TP391.9[自动化与计算机技术—计算机应用技术]