摘要
为了实现不同传感器影像之间的边缘特征匹配,提出了一种新的SAR影像边缘检测方法。该方法首先利用最小二乘支持向量(LS-SVM)对影像进行拟合以减小相干斑噪声的影响,然后采用基于小波变换的断面检测法进行影像边缘检测。通过与经典影像边缘提取算子的实验对比,表明该方法更适合于斑点噪声明显的SAR影像边缘检测。
A new edge detection method for synthetic aperture radar image is discussed. At first, fit the image with linear regress function based on least square support vector machine, which reduces the influence of speckle noise in image. Then, the direction of image edge is realized by wavelet transform and looking for local maximum values of gradient amplitude. The experiments show that the method could detect the image edge correctly.
出处
《测控技术》
CSCD
2007年第11期61-63,共3页
Measurement & Control Technology
基金
航天基金资助项目(0747-0540SITC2099-4)
关键词
合成孔径雷达影像
最小二乘支持向量机
小波变换
边缘检测
synthetic aperture radar image
least square support vector machine
wavelet transform
edge detection
