摘要
基于语义地图的自主导航移动机器人在空间勘探、危险物品搜索、自动驾驶等领域有着广阔的应用前景。为使移动机器人更好地完成相关复杂任务,针对传统SLAM(simultaneous localization and mapping)地图不够精细、语义信息缺失的问题,提出一种多传感器融合的语义地图构建技术。综合激光SLAM精度高和视觉信息纹理丰富的特点,选取基于ROS(robot operating system)系统的移动机器人平台,研究基于里程计、惯性传感单元(IMU)和激光雷达多源传感器融合进行即时定位与创建地图,将基于深度学习的目标检测算法移植到ROS系统中,结合机器人深度摄像头获取的图像信息实现目标语义识别,并结合深度信息进行位置解算实现目标定位和地图语义标注。通过机器人多目标语义添加实验和移动过程中的增量式地图构建与实时同步添加语义信息实验,验证该系统能实时地创建语义地图。
The mobile robot based on semantic map has broad application prospects in the fields of exploration,search of dangerous goods and automatic driving.Aiming at the problem that traditional SLAM does not have semantic information,in order to complete the complex tasks better,a semantic map construction technology based on multi-sensor fusion is proposed.The mobile robot platform based on ROS(robot operating system)is selected,and the real-time positioning and mapping based on the integration of odometer,inertial sensing unit(IMU)and lidar multi-source sensor is studied,and the target detection algorithm based on deep learning is transplanted.In the ROS,the target semantic recognition is realized by combining the image information acquired by the robot depth camera,and the target information and the map semantic annotation are realized by combining the depth information.Through the robot multi-target semantic addition experiment and the incremental map construction experiment,it is verified that the system can create a semantic map in real time.
作者
何松
孙静
郭乐江
陈梁
HE Song;SUN Jing;GUO Le-jiang;CHEN Liang(Air Force Early Warning Academy,Wuhan 430019,China;Air Force Communication Sergeant School,Dalian 116100,China;School of Computer Science,Wuhan University,Wuhan 430072,China)
出处
《计算机技术与发展》
2020年第9期88-94,共7页
Computer Technology and Development
基金
军委装发预研项目(315105305)。
关键词
移动机器人
同步定位与地图构建
语义信息
深度学习
目标检测
语义地图
mobile robot
simultaneous localization and mapping
semantic information
deep learning
target detection
semantic map
