中文摘要：

针对高分影像中过于冗余的细节信息和城区道路的复杂结构特征,提出一种基于结构线束统计匹配的城区主干道半自动提取方法。该方法由道路基线检测、线束剖面特征统计和道路模式匹配3部分构成,其核心是基于边缘统计特征的路段剖面结构表达。道路基线的确定使得剖面结构特征更加稳定;模式匹配阶段加入了道路几何结构的先验约束,从而抑制了提取结果的歧义性。实验选取了不同场景和范围下的城区高分遥感影像,定性和定量的实验分析结果表明,该方法在道路提取的精度和稳定性方面具有较好的表现,道路边界和中线的同时提取使其更具实用性。

英文摘要：

Aiming at the problems of redundant details on high resolution imagery and complex urban roads structure,we proposed a semi-automatic extraction method for urban roads based on line structure statistical matching.This method is composed of three parts：detecting the baseline of road,calculating statistical features of line profile,and road pattern matching.The core is the expression of road profile structure based on line statistical characteristics.Road baseline makes profile structure more stable.It suppresses the ambiguity of extraction results that priori constraints of road geometry are considered in pattern matching stage. High resolution remote sensing images of different urban scenes were taken as experimental data.The qualitative and quantitative analyses with experimental results show that this method has better performance in accuracy and stability of road extraction.It makes this method more practical that road boundary and center lines are extracted at the same time.