基于几何主轴和Hough变换的超宽带SAR二值边缘图像细化方法  

Thinning Algorithm Based on Geomatric Principle Axis and Hough Transform for UWB SAR Binary Edge Image

作　　者：焦春雨[1] 常文革[1] 

机构地区：[1]国防科技大学电子科学与工程学院,长沙410073

出　　处：《遥感信息》2013年第4期24-28,共5页Remote Sensing Information

基　　金：国家自然科学基金(61072116)

摘　　要：从经典形态学细化方法研究出发,结合超宽带SAR图像特点和超宽带SAR图像道路提取的应用背景,提出了一种基于几何主轴和Hough变换的超宽带SAR图像边缘细化方法。该方法克服了经典形态学细化方法边缘细化中容易出现的毛刺现象和单纯Hough变换对于区域方向判定的误差,提高了道路边缘细化的完整性。其细化后的线基元形态良好,有利于线基元连接成完整的道路。Thinning algorithm is a vital process in road extraction. Good thinning results can provide higher efficiency and accuracy in connecting the line segments. In this article, an efficient method of thinning binary edges of UWB SAR image, which is different from traditional morphlogical thinning algorithms,is presented. Combined the theory of geometric principle axis and Hough transform,the proposed method improves the integrity of the edges after thinning while avoiding the spurs and also de- clines errors while calculating direction of region in thinning. It also offers a good feature of line segments for line segment con- necting.

关 键 词：道路提取 二值图像细化 超宽带合成孔径雷达 

分 类 号：TP79[自动化与计算机技术—检测技术与自动化装置]