一种改进的LiDAR点云滤波方法  

An Improved LiDAR Point Cloud Filtering Method

作　　者：马小陆 甄文雪 巩朝光 MA Xiaolu;ZHEN Wenxue;GONG Chaoguang(School of Electrical and Information Engineering,Anhui University of Technology,Ma’anshan,Anhui 243000,China)

机构地区：[1]安徽工业大学电气与信息工程学院,安徽马鞍山243000

出　　处：《河北工程大学学报（自然科学版）》2022年第1期84-91,共8页Journal of Hebei University of Engineering:Natural Science Edition

基　　金：国家自然科学基金资助项目(61472282);安徽省科技重大专项(202003a05020028);安徽高校自然科学研究重点项目(KJ2019A0065)。

摘　　要：针对移动机器人激光雷达点云数据滤波过程中存在对噪声敏感和准确性差等问题,提出一种加权的高斯滤波方法。该方法使用高斯滤波与角度余弦作为加权值对雷达点云数据进行滤波处理,减小了激光雷达扫描到的物体轮廓波动。最后将该方法应用到实际的机器人操作系统的移动机器人底盘中,实验结果表明,相比传统的高斯滤波方法,该方法能更好地解决雷达点云滤波过程中的问题,扫描到的物体轮廓更加光滑,且具有较好的稳定性和准确性,有利于激光雷达建图效果。Aiming at solving the problems of poor noise sensitivity and accuracy in the filtering process of mobile robot LiDAR point cloud data,a weighted Gaussian filtering method was proposed.In this method,Gaussian filter and angle cosine were used as the weighted value to filter the radar point cloud data,which reduced the fluctuation of object contour scanned by LiDAR.Finally,the method was applied to the actual robot chassis of robot operation system.The experimental results show that the method can better solve the problems in the process of LiDAR point cloud filtering compared with the traditional Gaussian filtering method,and the scanned object profile is more smooth,and has good stability and accuracy,which is beneficial to the effect of LiDAR mapping.

关 键 词：点云 雷达建图 高斯滤波 角点提取 

分 类 号：TN958.98[电子电信—信号与信息处理]