摘要
通过对高分辨率遥感影像线状地物特征分析,提出了一种基于矩形模板匹配的高分辨率遥感影像线状地物提取方法.该方法由用户选取矩形种子区域,自动定义匹配区域,进行种子区域与匹配区域的最小二乘匹配,从而实现线状地物的半自动提取.与剖面匹配法和角度纹理法的对比实验研究结果表明,矩形模板匹配法能够快速、准确地提取出高分辨率遥感影像中的线状地物,节省了时间和人力.
After analyzing the features of linear objects in high resolution remote sensing images,based on rectangular template matching,a method for extraction of linear features in high resolution remote sensing images is presented in this paper.With this method,the user can choose the seed rectangle region,and the system can automatically define the matching region and complete the least squares matching of the seed region and the matching region,thus realizing a semi-automatic extraction of linear features.To verify this method,we did a contrast experiment with the methods of rectangular template matching,profile matching and angle texture matching.The research and the experiment results showed that this method can extract linear features in high resolution remote sensing images rapidly and accurately.It is of good practical value,for it can save time and human labor.
出处
《西南大学学报（自然科学版）》
CAS
CSCD
北大核心
2015年第7期155-160,共6页
Journal of Southwest University(Natural Science Edition)
基金
国家自然科学基金(41301417)
国家发改委卫星及应用产业发展专项([2013]2140号)
关键词
矩形模板
最小二乘匹配法
线状地物
半自动提取
高分辨率遥感影像
rectangular template
least squares matching method
linear feature
semi-automatic extraction
high resolution remote sensing image
