摘要
正射影像镶嵌过程中,镶嵌线的选取是一个重要的步骤,其自动化程度是影响全自动镶嵌的一个重要因素。本文提出一种基于图论最短路径原理的正射影像镶嵌线自动提取方法。该算法将图像上提取得到的底层视觉信息(Canny边缘)作为镶嵌过程中的障碍物信息,在边缘图像上根据像素点的邻接关系构建带权有向图,利用初始镶嵌线作为初始条件计算图中有向边的权值,将正射影像镶嵌线的提取过程转化为图论中最短路径问题。实验表明,该算法可以较为准确提取镶嵌线,对全自动镶嵌具有重要应用价值。
Seamlines extractions are important step during DOM mosaicking,which is an important factor for automatic processing.An automatic seamline extraction algorithm for orthophoto mosaicking based on shortest path method is introduced. The algorithm regards low-level vision information (Canny edge)as obstacles information during mosaicking processing, constructs weighted directed graph using Canny edge image according pixels’neighbors,calculates weight for directed edge in graph by initial seamline,and converts seamline extraction for orthophoto mosaicking to shortest path problem.Result shows that the algorithm can extract seamline precisely,and is useful for automatic DOM mosaicking processing.
出处
《遥感信息》
CSCD
北大核心
2015年第1期33-36,共4页
Remote Sensing Information
基金
国家科技支撑计划课题(2011BAB01B05)
国家自然科学基金项目(41101435)
关键词
正射影像
镶嵌线
带权有向图
最短路径
digital orthophoto map
seamline
weighted directed graph
shortest path
