摘要
针对传统基于标志物的增强现实跟踪注册方法的使用场景限制问题,提出改进的ORB-SLAM系统。论文采用SURF-ORB算法,对环境特征点进行提取匹配,构建真实环境的地图库,对运动物体进行定位。使用改进后的方法,提高了特征点检测和匹配的速度以及定位跟踪的稳定性。
An improved ORB-SLAM system is proposed to overcome the problem of the use of traditional marker-based re?al-time tracking method.In this paper,a SURF-ORB algorithm is proposed to extract and match the environmental feature points,a map library of real environment is constructed,and the moving objects are located.Using the improved method,the speed of fea?ture detection and matching and the stability of positioning and tracking are improved.
作者
闫兴亚
范瑶
YAN Xingya;FAN Yao(School of Computing,Xi'an University of Posts&Telecommunications,Xi'an 710061)
出处
《计算机与数字工程》
2018年第12期2565-2569,共5页
Computer & Digital Engineering
基金
西安市科学技术局"基于TLD的增强现实跟踪注册方法在电梯维修系统中的应用实现"(编号:2017084CG/RC047(XAYD002))
西安市科学技术局"基于unity3D的低延迟VR交互设计在电梯安装中的应用"(编号:2017084CG/RC047(XAYD003))资助