多站激光扫描点云配准算法研究  

Research on Point Clouds Registration Algorithm by Multiple Scans

作　　者：刘志超 顾广杰 LIU Zhichao;GU Guangjie(61206 Troops,Dalian 116023,China)

机构地区：[1]61206部队,辽宁大连116023

出　　处：《海洋测绘》2018年第5期36-40,共5页Hydrographic Surveying and Charting

摘　　要：随着Li DAR的迅速发展,三维激光扫描技术已经被广泛运用于各个领域。点云数据配准过程中,传统的配准算法比较依赖特征点匹配精度,粗差点的存在会较大程度地影响配准精度和配准效率。通过对罗德里格矩阵、整体最小二乘原理的分析,提出了一种基于整体最小二乘的罗德里格矩阵算法。在该算法实现的过程中,能够考虑到系数矩阵误差,降低算法实现过程中特征点坐标误差对参数求解的影响。实验结果表明,本文算法比参数算法和严密的罗德里格算法精度更高、稳定性更强,在初始对应点坐标误差较大的情况下仍能获得精度较为稳定的变换参数。With the rapid development of LiDAR, 3D laser scanning technology has been widely used in many fields.In the process of the Lidar point cloud data processing, some traditional algorithms rely more on the matching accuracy of feature points, so that the presence of coarse points will greatly affect the registration accuracy and efficiency.Therefore,through the analysis of Lodrigues Matrix and Total Least Squares principle,a Lodrigues Matrix algorithm based on Total Least Squares principle is proposed. In the implementation of the algorithm, the error of the coefficient matrix can be taken into account, and the influence of the coordinate error of characteristic points in the algorithm realization is reduced.Finally, the experiments showed that this algorithm has a higher accuracy and robustness, which maintains stable transformation parameters while there exit biggish errors in the initial coordinates of corresponding points.

关 键 词：LIDAR 点云配准 刚体变换 罗德里格算法 整体最小二乘 

分 类 号：P231[天文地球—摄影测量与遥感]