提高三维激光扫描点云数据配准精度研究  

Research on the Improvement of Registration Accuracy of Three-dimensional Laser Scanning Point Clouds

作　　者：于明旭[1] YU Mingxu(Jiangsu Vocational Institute of Architectural Technology,Xuzhou 221116,China)

机构地区：[1]江苏建筑职业技术学院,江苏徐州221116

出　　处：《测绘与空间地理信息》2020年第5期47-50,55,共5页Geomatics & Spatial Information Technology

基　　金：2018年度江苏省建设系统科技项目(指导类)(2018ZD009);2017年江苏建筑职业技术学院科研课题(JYA317-12)资助。

摘　　要：点云数据配准是三维激光扫描数据处理中的关键步骤之一,本文结合工程实例,运用不同方法进行点云初始配准,初始配准结果作为ICP算法的初始位姿进行精配准,实验证明,初始位姿对点云的配准精度、速度有一定影响。在工程实践中,提高初始配准精度,是解决ICP算法易陷入局部最优解问题及ICP算法精确配准效率低下问题的有效途径之一。Point cloud data registration is one of the key steps in 3D laser scanning data processing.The paper uses different methods to carry out initial point cloud registration.The initial registration results provide ICP algorithm with initial position for precise registration.The experiment proves that the initial position has certain influences on the registration accuracy and speed.To improve the initial registration accuracy is one of the effective ways to solve the problem that ICP algorithm is easy to get stuck at the local optimal solution and its low efficiency.

关 键 词：点云数据 初始配准 精配准 ICP算法 三维激光扫描 

分 类 号：P225.2[天文地球—大地测量学与测量工程]