基于点云预处理的路面三维重构数据优化  

Data optimization of pavement 3D reconstruction based on point preprocessing

作　　者：魏亚[1] 肖庸 闫闯 刘亚林 汪林兵 WEI Ya;XIAO Yong;YAN Chuang;LIU Ya-lin;WANG Lin-bing(School of Civil Engineering,Tsinghua University,Beijing 100084,China;National Center for Materials Service Safety,University of Science and Technology Beijing,Beijing 100083,China)

机构地区：[1]清华大学土木水利学院,北京100084 [2]北京科技大学国家材料服役安全科学中心,北京100083

出　　处：《吉林大学学报（工学版）》2020年第3期987-997,共11页Journal of Jilin University:Engineering and Technology Edition

基　　金：国家重点研发计划项目(2017YFF0205600).

摘　　要：提出了路面点云预处理方法,改善了常规方法在去噪、平滑和采样阶段存在的问题。首先,使用基于局部密度的去噪算法对点云零散噪声进行处理,消除了常规去噪方法适应性差或复杂度高的问题。然后,针对去噪后的点云,采用基于移动最小二乘法的平滑和采样连续算法,在区分路面特征区与非特征区的基础上,改善了平滑与采样阶段割裂造成的数据二次劣化问题。最后,进行了实际路面试验验证,结果表明:改进后的去噪、平滑和采样方法提高了建模精度,优化了三维扫描所获数据,提供了一种新的基于点云预处理的重构方法。A new preprocessing method of pavement point cloud is proposed,which solves the problems existing in the procedure of outlier removal,smoothing and sampling of conventional methods.First,a local density-based outlier removal algorithm is used to deal with the scattered noise of the point cloud,which solves the problems of poor adaptability or high complexity of conventional methods.Then for the denoised point clouds,a continuous algorithm based on the moving least squares method is adopted to smooth and sample the point.Finally,on the basis of distinguishing the featured and non-featured areas of pavement,the problem of data repeated degradation caused by the splitting of smoothing and sampling stages is solved.The improved outlier removal,smoothing and sampling methods have been verified by actual pavement tests to improve the accuracy of modeling,which can optimize the data obtained by threedimensional laser scanning,and provide a new reconstruction method based on the point cloud preprocessing.

关 键 词：道路与铁路工程 局部密度 平滑 去噪 移动最小二乘法 

分 类 号：U416[交通运输工程—道路与铁道工程] TP391[自动化与计算机技术—计算机应用技术]