摘要
针对合成孔径雷达 (SAR ,SyntheticApertureRadar)图像的特点 ,提出一种线特征提取算法 .该算法首先抑制相干斑噪声 ,然后对SAR图像进行模糊变换和增强 ,再基于假设的SAR图像灰度模型 ,用最小二乘法设计出低通和高通滤波器 ,最后检测出线特征散点 ,并进行连接 ,得到SAR图像的线特征图 .实验证明本算法抗噪声干扰 ,边缘定位精度达到 1个像素 .
A linear feature extraction method is proposed for SAR image. Firstly, the speckle noise in the radar image is filtered out. Secondly, the SAR image is processed by fuzzy transition and enhancement. Then, the low-pass and high-pass filters are designed by the least squares methods and the gradient strength map and direction map are acquired. Lastly, the gradient profile maximum method is adopted to detect feature points, which are connected to form the linear feature map of the SAR image. Simulation results show that the proposed method has the characteristic of anti-noise, and the precision of positioning can reach 1 pixel.
出处
《北京航空航天大学学报》
EI
CAS
CSCD
北大核心
2003年第4期342-345,共4页
Journal of Beijing University of Aeronautics and Astronautics
基金
航空基础科学基金资助项目 ( 99E5 10 18)
关键词
合成孔径雷达
特征抽取
最小二乘法
线特征
模糊增强
Fuzzy sets
Image processing
Least squares approximations
Synthetic aperture radar
