面向室内环境的主动视觉即时定位与地图构建系统

(1.哈尔滨理工大学自动化学院,黑龙江哈尔滨150080;2.中国科学院自动化研究所,北京100190)

计算机视觉;障碍物检测;避障;行人检测;地图构建

An active visual simultaneous localization and mapping system for mobile robot in indoor environment
WANG Hongmin1* ,SUN Zengpeng1,ZHAO Xiaoguang2,BIAN Jiang2

(1.School of Automation,Harbin University of Science and Technology,Harbin 150080,China;2.Institute of Automation,Chinese Academy of Sciences,Beijing 100190,China)

computer vision;obstacles detection;obstacles avoidance;pedestrians detection;mapping

DOI: 10.6043/j.issn.0438-0479.201810008

备注

提出了一个主动视觉即时定位与地图构建(SLAM)系统,能够避开障碍物,预测行人运动方向并且躲避行人,同时能够在未知环境中获得有效路径和进行稀疏三维点云地图构建.该系统由机器人平台、RGB-D摄像机和双目摄像机构成.RGB-D摄像机基于RGB-D上半身探测器进行行人检测与跟踪,同时使用RGB-D摄像机进行三维地图构建.双目摄像机通过获取深度信息寻找可通行路径.该系统实现了主动地图构建并且避免了由传统方法构建的静态地图包含行人的情况.实验验证了该系统的有效性.

We present an active visual Simultaneous Localization and Mapping (SLAM) system which is capable of avoiding obstacles,predicting the direction of pedestrians and avoiding pedestrians.Simultaneously,SLAM can obtain effective paths and 3-D map in unknown environments.It consists of a robot platform,an RGB-D camera and a stereo camera.The RGB-D camera is used for 3-D mapping and pedestrian detection as well as tracking by RGB-D upper-body detectors.The stereo camera finds a passable path by leveraging the depth information.The proposed system enables active mapping and avoids the typical situations,in which people will