一种SAR影像带状道路半自动提取方法  

A semi-automatic extraction of ribbon roads from SAR image

作　　者：孙娜[1,2] 张继贤[1] 黄国满[1] 赵争[1] 

机构地区：[1]中国测绘科学研究院,北京100830 [2]辽宁工程技术大学,辽宁阜新123000

出　　处：《测绘科学》2014年第11期108-113,14,共7页Science of Surveying and Mapping

基　　金：863项目(2011AA120401)

摘　　要：文章提出了一种结合极化的相干最优信息的SAR影像带状道路半自动提取方法。该方法由用户输入种子点确定初始矩形模板,用傅立叶函数构建道路轨迹模型作为全局的控制条件,将矩形模板的角度信息结合极化相干最优信息作为局部的控制条件,根据两种条件控制模板的前进方向,实现道路的边缘线提取。通过与人工提取结果对比,证明本文的方法能够取得较理想的效果。This paper proposed a semi-automatic extraction method of ribbon roads from SAR image combined with polarized coherence optimization information. By this method, users input seed point to de- termine the initial rectangular template, and it builds road trajectory model as a global control condition u- sing Fourier function, makes the rectangular template angular information combined with polarized coher- ence optimization information as a local control condition, and then extracts the road edge lines according to the two control conditions. By comparing with the manual extraction, it was proved that the method could obtain satisfactory results.

关 键 词：SAR影像 半自动道路提取 极化相干最优信息 角度信息 道路轨迹模型 

分 类 号：TP751[自动化与计算机技术—检测技术与自动化装置] P236[自动化与计算机技术—控制科学与工程]