变换域SAR图像边缘检测  

Edge detection in SAR images of transform domain

作　　者：梅珍[1] 林伟[1] 王瑞霞[1] 

机构地区：[1]西北工业大学理学院,西安710129

出　　处：《计算机应用》2009年第9期2402-2405,共4页journal of Computer Applications

摘　　要：针对合成孔径雷达(SAR)图像含有大量相干斑噪声,采用变换域中多尺度乘积进行边缘提取。该方法利用信噪比最大自动选取阈值,在非下采样Contourlet变换(NSCT)域中,进行不同方向高频系数多尺度乘积以达到SAR图像检测边缘。实验结果显示,与传统的边缘检测方法相比,在非下采样Contourlet变换域中多尺度积的信噪比最大的边缘检测方法得到更好的结果。Concerning plenty of speckle noise in Synthetic Aperture Radar （SAR） images, muhiscale products method was proposed to detect edge in transform domain. This method was to obtain the value of a threshold automatically using the max signal-to-noise ratio and made different directions multiscale products of high frequency coefficient in order to detect edge of SAR images in NonSubsampled Contourlet Transform （NSCT） domain. Compared with classical methods, the experimental results demonstrate the validity of the proposed method.

关 键 词：非下采样CONTOURLET变换 合成孔径雷达图像 多尺度积 信噪比 边缘检测 

分 类 号：TP951[自动化与计算机技术] TP391.41