摘要
基于整体矩形匹配方法、结点扩展方法、道路连续匹配终止准则等,提出了一种较完善的从高分辨率遥感影像上提取直线道路方法.利用最佳匹配矩形扩展结点进行道路的连续提取,建立道路连续匹配终止准则结束对当前道路的连续提取.该方法具有较高的自动化程度,能够快速准确地提取出影像上主要道路,较好地表示道路的边缘、宽度和方向.同时,在道路提取过程中将自动跟踪和少量人工处理有机地结合起来,使得方法具有实用性.
A more perfect road extraction approach from high-resolution remote sensing image is proposed. In this extraction, the extending node method extracted roads continuously through total rectangle matching algorithm; the extraction road was stopped continuously by the ending rule when the extraction needed ending. Thus, this method achieves the extraction of main roads with high automatically and accurately. The edges, widths and directions of roads can be represented preferably by the approach. A little manual processing is incorporated in auto-tracking to make the method more practical.
出处
《华中科技大学学报（自然科学版）》
EI
CAS
CSCD
北大核心
2009年第1期27-29,48,共4页
Journal of Huazhong University of Science and Technology(Natural Science Edition)
基金
国家科技支撑计划资助项目(2007BAH16B03)
关键词
遥感影像
道路提取
自动化
整体矩形匹配
结点扩展
终止准则
remote sensing image
road extraction
automation
total rectangle matching
extendingnode
ending rule
