新型Canny算法的SAR图像边缘检测  

A new method of edge detection based on improved Canny operator for SAR image

作　　者：刘景正[1] 余旭初[1] 张鹏强[1] 杨国鹏[1] 

机构地区：[1]解放军信息工程大学测绘学院

出　　处：《测绘科学》2008年第1期132-134,共3页Science of Surveying and Mapping

摘　　要：干涉作用会使合成孔径雷达回波信号产生相干斑点噪声,以致在图像边缘检测时产生许多虚假边缘;Canny边缘检测器能检测图像形状密度的变化,并寻找图像的边界。针对SAR图像的噪声特点和低像素级的特点,引进了小波去噪的方法来代替传统的高斯滤波,提出了在搜索算法上以广度搜索来代替深度搜索,由此改进了Canny算法,使用VC++编程实现了改进的算法对SAR图像进行边缘提取实验,对比结果证明此方法比以往算法有明显改进。The coherent signals sent from SAR will interfere with the images receipted from objects, and could result in speckles which produces lots of edge shadows during edge detection.The Canny edge detection operator can check and measure the change in shape and density,and find out the edge of the image.Focusing on the characteristic of speckle and low pixels of SAR image, this thesis introduces the Wavelet speckle reduction and breadth-first traversal instead of the traditional depth-first search into Canny operator.With VC++ programming, it becomes possible to do the SAR image edge detection experiment by improved algorithm.The result proves its advantage over the old method.

关 键 词：CANNY 边缘检测 SAR 噪声 

分 类 号：TP751[自动化与计算机技术—检测技术与自动化装置]