摘要
提出一种适用于影像匹配粗差剔除的局部矢量面元方法。首先消除匹配像对间的系统性差异;然后通过构建匹配同名点的三角网结构,实现匹配结果的局部面元分割;在局部面元上进行矢量统计,引入针对局部敏感的矢量描述子指标,根据误差分布满足正态分布规律的假设设定合理的阈值,并最终实现影像匹配粗差的快速剔除。两组数据的试验验证了所提方法的可行性,算法处理速度快、剔除成功率高。
This paper proposes a method that can be applied to eliminating gross errors in image matching.The whole process can be divided into three steps.Firstly,the systematic difference is removed.Then triangulated irregular network(TIN)of image matching points is constructed to get the partitioning local field.Based on the normal distribution of the image matching gross error,a vector descriptor is proposed in the statistics on the local field.Finally,a reasonable threshold is used in eliminating gross errors.The feasibility of the proposed method is verified based on the experiments of two groups of data.The results showed high processing speed and success rate of gross error elimination.This method also provided a new viewpoint to the research of photographic error processing and reliability theory.
出处
《测绘学报》
EI
CSCD
北大核心
2014年第7期717-723,共7页
Acta Geodaetica et Cartographica Sinica
基金
国家自然科学基金(41322010
41171292)
国家863计划(2013AA12A401)
教育部博士研究生学术新人奖(5052012213002)
武汉大学研究生自主科研项目(2012213020201)
关键词
影像匹配
粗差剔除
局部面元
矢量描述子
image matching
gross error elimination
local field
vector descriptor
