基于视觉SLAM的煤矿救援机器人建图研究

Research on coal mine rescue robot based on visual SLAM

当煤矿事故发生之后,为了能够快速进行灾后勘探和救援任务,利用开源的ORB-SLAM2算法和深度相机,结合已有的移动机器人平台,对位于井下的移动机器人的定位以及巷道环境的建图进行研究。分别在光线充足和有辅助光源的黑暗条件下进行模拟试验,得出了移动机器人的运动轨迹,为实际井下救援过程中煤矿救援机器人的自主导航系统提供尺度准确的2D栅格地图。

After the coal mine accident occurred,in order to quickly carry out post-disaster exploration and rescue tasks,the open source ORB-SLAM2 algorithm and depth camera were used,combined with the existing mobile robot platform,which were used to study the positioning of the underground mobile robot and the construction of the tunnel environment.The simulation experiments were carried out under the dark conditions of sufficient light and auxiliary light source,and the motion trajectory and color original point cloud map were obtained,which provided an accurate 2D grid map for the autonomous navigation system of coal mine rescue robots in the actual underground rescue process.