摘要
针对随机抽样一致算法在误匹配剔除时存在稳定性不足、效率较低等问题,提出一种粗剔除与精剔除相结合的误匹配消除算法.该算法首先利用最小距离法对特征点进行筛选获得初始匹配点集;然后通过计算特征点的相关性实现精剔除;最后将该算法应用于ORB-SLAM2系统进行验证.试验结果表明,该算法可有效剔除误匹配特征点,获得匹配精度更高的匹配点集,在ORB-SLAM2系统中能较好地获得相机轨迹,运行效果佳.
To solve problems of insufficient stability and low efficiency in random sampling consensus(RANSAC)algorithm when mismatching is eliminated,a mismatch elimination algorithm that combines coarse elimination and fine elimination is proposed.Minimum distance method is firstly used to screen feature points for obtaining initial matching point set,and then correlation of the feature points is calculated for precise elimination.Finally,the algorithm is used to apply to ORB-SLAM2 system.The experimental results show that the algorithm can effectively eliminate mismatched feature points and obtain a set of matching points with higher matching accuracy.In ORB-SLAM2 system,camera trajectory can be obtained easily with remarkable effect.
作者
张盟
王志亮
刘汉忠
ZHANG Meng;WANG Zhi-liang;LIU Han-zhong(School of Mechanical Engineering,Nanjing Institute of Technology,Nanjing 211167,China;School of Automation,Nanjing Institute of Technology,Nanjing 211167,China)
出处
《南京工程学院学报(自然科学版)》
2021年第1期24-28,共5页
Journal of Nanjing Institute of Technology(Natural Science Edition)
基金
国家自然科学基金项目(51675259)。