摘要
针对需要具有旋转不变性且具有实时性的任务场景,传统的局部特征提取算法SIFT与SURF不能满足实时性要求的现状,提出了一种基于ORB特征检测子的优化特征点匹配算法。首先针对在原始ORB特征匹配算法中出现的错误匹配问题,利用特征点的位置信息结合聚类算法提高匹配过程的速度与正确率,再通过均值漂移算法进一步提取出错误匹配点对。将所提方法应用于生产线产品外观缺陷检测设备,经过实际实验验证,该算法在ORB特征匹配中正确率提高至95%,能够满足实时使用的需要。
As traditional SIFT(Scale-Invariant Feature Transform)and SURF(Speeded Up Robust Features)local feature points matching algorithm dissatisfy real-time jobs,an optimized feature point matching algorithm based on ORB(Oriented FAST(Features from Accelerated Segment Test)and Rotated BRIEF(Binary Robust Independent Elementary Features))feature detection was proposed.For mismatched pairs in the original ORB feature matching algorithm,the position information of feature points was combined with the clustering algorithm to promote the speed and correct rate in the matching process,and then the mismatched pairs were further extracted by the mean shift algorithm.The proposed algorithm was applied to the appearance defect detection equipment of the production line.In the actual experiment,the accuracy of the proposed algorithm in ORB feature matching was improved to 95%,which can satisfy the needs of real-time jobs.
作者
杨溪远
陈斌
YANG Xiyuan;CHEN Bin(Chengdu Institute of Computer Application,Chinese Academy of Sciences,Chengdu Sichuan 610041,China;University of Chinese Academy Sciences,Beijing 100049,China;Guangzhou Electronic Technology of Chinese Academy of Sciences,Guangzhou Guangdong 510070,China)
出处
《计算机应用》
CSCD
北大核心
2019年第S02期81-84,共4页
journal of Computer Applications
基金
广东省重大科技专项(2017B03030617)
广东省产学研合作项目(2017B090901040)
关键词
ORB特征
特征点匹配
均值漂移
局部特征
图像对准
ORB(Oriented FAST(Features from Accelerated Segment Test)and Rotated BRIEF(Binary Robust Independent Elementary Features))feature
feature point matching
mean shift
local feature
image alignment
