基于共面条件的点云配准旋转角参数求解方法  

A Registration Algorithm for Point Cloud Rotation Angle Based on Planar Unit Normal Vector

作　　者：朱宁宁[1] 

机构地区：[1]河南理工大学测绘与国土信息工程学院,河南焦作454003

出　　处：《测绘与空间地理信息》2015年第5期202-205,210,共5页Geomatics & Spatial Information Technology

摘　　要：针对目前Li DAR获取的点云数据配准中ICP算法迭代计算效率较低和使用标靶配准时精扫标靶费时费力等问题,提出利用扫描地物所包含的平面特征及点云数据的离散特性,通过拟合平面得到平面的单位法向量进行旋转角的求解。由于点云数据中含有误差,使用拟合平面的法向量不仅避免了对标靶的精细扫描,而且也消弱了点云误差对转换参数的影响。最后通过实例验证了本文方法的可行性与严密性。In view of the present LiDAR to obtain the point cloud data registration, the ICP algorithm iterative calculation efficiency is low and target with fine scanning consuming much time, put forward take the advantage of plane characteristics of scanning object and the discrete characteristics of point cloud data, the unit normal vector of plane is obtained by fitting plane and through this got the rotation angle. At last, the feasibility and rigor of the presented approach are verified by examples.

关 键 词：LIDAR 平面拟合 单位法向量 点云配准 旋转角点云 

分 类 号：P25[天文地球—测绘科学与技术] TP303[自动化与计算机技术—计算机系统结构]