摘要
同时定位与地图构建(simultaneous localization and mapping,SLAM)是自主移动机器人导航的一个重要研究方向。视觉SLAM是一种使用相机作为主要信息来源的SLAM技术。与较为成熟的基于激光测距的SLAM相比,视觉SLAM还有许多问题亟待研究解决。近年来,随着人工智能、机器学习、图像处理技术、云计算、5G通信技术的发展,计算机视觉的相关算法能够在终端设备实时运行,视觉SLAM研究也受到越来越多的关注。本文介绍了视觉SLAM的原理及一些经典的实现方法,概述了视觉SLAM的最新研究进展,对视觉SLAM在实际应用中存在的一些问题进行了讨论。
Simultaneous localization and mapping(SLAM)was an important research direction for autonomous mobile robot navigation.Visual SLAM was a SLAM technology that used a camera as the main source of information.Compared with the more mature SLAM based on laser ranging,visual SLAM still had many problems to be solved urgently.In recent years,with the development of artificial intelligence,machine learning,image processing technology,cloud computing,and 5G communication technology,the related algorithms of computer vision could run in real-time on terminal devices,and visual SLAM research also attracted more attention.The principles of visual SLAM and some classic implementation methods were introduced,the latest development of visual SLAM was summarized,and some problems in the practical application of visual SLAM were discussed.
作者
张大伟
苏帅
ZHANG Dawei;SU Shuai(School of Information Engineering,Zhengzhou University,Zhengzhou 450001,China;School of Electronics and Information Engineering,Tongji University, Shanghai 201804,China)
出处
《郑州大学学报（理学版）》
CAS
北大核心
2021年第1期1-8,共8页
Journal of Zhengzhou University:Natural Science Edition
基金
国家自然科学基金项目(61903341)
河南省科技攻关项目(202102210133)。
