井下障碍物激光雷达动态识别技术

针对井下行人车辆、支护设施、管网线缆等环境障碍物严重影响井下无人机飞行安全的难题，防范飞行故障甚至设备损毁，提出一种井下障碍物激光雷达动态识别方法：采用机载激光雷达对井下环境进行在线扫描并获得高分辨率的点云数据，基于点云滤波预处理、随机采样一致性(RANSAC)点云分割、欧氏聚类点云聚类方法和主成分分析(PCA)包围盒方法，联合实现对井下环境障碍物的有效实时识别。该方法障碍物识别率达85%,可有效保障井下无人机飞行探测作业安全，服务于智能矿山建设和矿山安全救援需求。

Research on underground obstacle dynamic detection based on LiDAR

Aiming at the problem that environmental obstacles such as underground pedestrian vehicles, support facilities, pipe network and cables seriously affect the flight safety of underground UAV, in order to prevent flight faults and equipment damage, this paper proposes a dynamic identification method of underground obstacle detection, which uses airborne LiDAR to scan the underground environment online and obtain high-resolution point cloud data. The method jointly realize the effective real-time identification of underground environmental obstacles through the preprocessing method based on point cloud filtering, the point cloud segmentation method based on random sampling consistency (RANSAC), the point cloud clustering method based on Euclidean clustering and the bounding box method based on principal component analysis (PCA). The obstacle recognition rate is up to 85%, which effectively ensures the safety of underground UAV flight detection operation and can better serve the construction of Intelligent Mine and mine safety rescue.