无人水面艇三维激光雷达目标实时识别系统  

3D LiDAR Based Real-time Object Recognition System for Unmanned Surface Vehicles

作　　者：柳晨光[1,2] 郭珏菡 吴勇[1,2,3] 初秀民[1,2] 吴文祥 雷超凡 LIU Chenguang;GUO Juehan;WU Yong;CHU Xiumin;WU Wenxiang;LEI Chaofan(National Engineering Research Center for Water Transport Safety,Wuhan University of Technology,Wuhan 430063;Intelligent Transport System Research Center,Wuhan University of Technology,Wuhan 430063;School of Transportation and Logistics Engineering,Wuhan University of Technology,Wuhan 430063)

机构地区：[1]武汉理工大学国家水运安全工程技术研究中心,武汉430063 [2]武汉理工大学智能交通系统研究中心,武汉430063 [3]武汉理工大学交通与物流工程学院,武汉430063

出　　处：《机械工程学报》2022年第4期202-211,共10页Journal of Mechanical Engineering

基　　金：国家自然科学基金(52001240);湖北省自然科学基金(2020CFB307);重庆市自然科学基金(cstc2021jcyj-msxm X1220);中央高校基本科研业务费专项资金资(203144003,202444001,213244001)资助项目。

摘　　要：为解决无人水面艇动态环境目标动态感知问题,研究无人艇三维激光雷达目标实时识别系统。设计出无人艇三维激光雷达目标实时识别系统结构、硬件组成及数据通信协议。基于点云库(Point cloud library,PCL)、Qt和Visual Studio平台开发了无人艇三维激光雷达目标实时识别系统软件,实现了点云数据校正、实时处理、数据显示、状态输出、远程通信等功能。考虑到无人艇航行时周边环境障碍物三维激光点云分布特征,将三维激光点云投影至多属性二维栅格进行表示,利用八邻域算法实现了障碍物栅格的聚类,解决了点云数据处理、目标分割、点云图像远程交互等关键技术。最后,构建了室外水池环境下的无人艇三维激光雷达目标实时识别系统试验平台,测试结果表明该系统能够可靠、准确识别无人艇周围100 m范围内的障碍目标。In order to solve the problem of unmanned surface vehicles(USVs) dynamic environment target dynamic perception, a 3D(three dimension) LiDAR based real-time object recognition system is studied. The structure, hardware components and data communication protocol of the 3D LiDAR based real-time object recognition system are designed. Based on the point cloud library(PCL) library, Qt and Visual Studio platform, the system software is developed, which realizes point cloud data correction, real-time processing, data display, status output, and remote communication and other functions. Considering the characteristics of the 3D point cloud in the surrounding environment of a sailing USV, the 3D laser point cloud is projected to a multi-attribute two-dimensional grid for representation, and the eight-neighbor algorithm is used to realize the clustering of the obstacle grid a. Key technologies such as point cloud data processing, target segmentation, and remote interaction of point cloud images are solved. Finally, the platform of 3D Li DAR based real-time object recognition system for a USV in an outdoor pool environment is constructed, and the results show that the system can reliably and correctly identify the obstacles within a range of 100 m around the USV.

关 键 词：无人艇 激光雷达 三维点云 目标识别 栅格地图 八邻域 

分 类 号：U675[交通运输工程—船舶及航道工程]