摘要
为提高不同视角图像的匹配精度,提出一种多视角图像的图变换匹配算法。利用传统图像匹配算法获取多视角图像特征点的初始匹配关系;计算特征点之间的欧氏距离,建立近邻图判断初始匹配的正确性;对错误的特征匹配关系进行剔除,提高图像的匹配精度。多组不同场景的多视角图像实验结果表明,与随机采样一致性算法(random sample consensus,RANSAC)、图变换匹配算法以及迭代的图变换匹配算法进行对比,该算法在正确匹配点对的查准率及错误匹配点对的查全率上均取得较好结果,验证了其在多视角图像匹配上的有效性。
To improve the matching accuracy of different view images,a graph transformation matching algorithm for multi-view images was proposed Initial feature point matching was got using the traditional image matching.The Euclidean distance be-tween feature points was calculated.Nearest neighbor graphs were built to determine whether the initial match was correct.The feature points matching errors were eliminated,so as to improve the matching accuracy of the multi-view images.The algorithm was compared with the random sample consensus algorithm,graph transformation matching algorithm and the iterative graph transformation matching algorithm(RANSAC)through multi group experiments.The results show that the proposed algorithm can improve precision ratio of the correct matching points and the recall ratio of wrong matching points,which verifies the effec-tiveness of the algorithm in multi-view images matching.
作者
温佩芝
成龙
龚震霆
赵萌
WEN Pei-zhi;CHENG Long;GONG Zhen-ting;ZHAO Meng(School of Computer Science and Information Security, Guilin University of Electronic Technology, Guilin 541004, China;School of Electronic Engineering and Automation, Guilin University of Electronic Technology, Guilin 541004,China;Guangxi Colleges and Universities Key Laboratory of Intelligent Processing of Computer Images and Graphics,Guilin University of Electronic Technology, Guilin 541004,China)
出处
《计算机工程与设计》
北大核心
2017年第2期442-448,共7页
Computer Engineering and Design
基金
国家自然科学基金项目(61063019)
广西自然科学基金项目(桂科自0991240)
广西科技计划项目基于图像的古建筑立面及构件测量方法研究基金项目(桂科攻14124005-2-9)
桂林电子科技大学研究生创新基金项目(GDYCSZ201418)
广西高校图像图形智能处理重点实验室基金项目(LD15043X)
关键词
多视角图像
图变换匹配
近邻图
查准率
查全率
multi-view images
graph transformation matching
nearest neighbor graph
precision ratio
recall ratio
