摘要
同步定位与构图(simultaneous localization and mapping,SLAM)使机器人可快速构建场景地图并确定自身位置,是室内定位与导航的基本模块。依此设计实现一种使用2D激光传感器进行室内数据采集的3D激光SLAM系统,具有成本低、易携带、可拓展等特点。该系统基于室内平面和相邻面交线为特征,利用鲁棒的特征匹配进行特征关联,通过分级位姿优化策略,逐步优化位姿轨迹,解决了低成本采集设备数据量小、频率低、帧内变形大的问题,实现低成本、低漂移的室内激光SLAM技术。最后通过定性定量的对比实验验证该系统模块的有效性和普适性。
Simultaneous localization and mapping(SLAM)enables the robot to quickly build a scene map and determine its own position,which is a basic module for indoor positioning and navigation.This paper designs and implements a 3D laser SLAM system that uses the motor to drive the 2D laser sensor to perform indoor data collection,which has the characteristics of low cost,portability,and expansibility.This algorithm features the widely existing plane and adjacent plane intersections,applies robust feature matching to perform feature correlation,and adopts a hierarchical pose optimization strategy to gradually optimize the pose trajectory.Finally,qualitative and quantitative comparison experiments are performed to verify the robustness of feature expression and matching,the effectiveness of the hierarchical pose optimization strategy,and universality.
作者
罗磊
谈彬
姚剑
李礼
LUO Lei;TAN Bin;YAO Jian;LI Li(School of Remote Sensing and Information Enfineering,Wuhan University,Wuhan 430079,China)
出处
《测绘地理信息》
CSCD
2022年第4期56-60,共5页
Journal of Geomatics
基金
国家重点研发计划(2017YFB1302400)
关键词
同步定位与构图
激光传感器
室内定位
相邻面交线
激光里程计
SLAM
laser sensor
indoor positioning
adjacent plane intersections
LOAM(LiDAR odometry and mapping)
作者简介
第一作者:罗磊,硕士生,主要从事激光SLAM技术研究。E-mail:lei.luo@whu.edu.cn;通讯作者:姚剑,博士,教授,主要从事三维立体技术研究。E-mail:jian.yao@whu.edu.cn
