摘要
地面三维激光点云配准是点云分析、三维重建以及其他工程应用的基础和关键技术。针对现有配准算法精度和效率上的不足,提出了一种基于快速点特征直方图及相邻两站间重叠区计算的配准方法。粗配准阶段,在经典的点特征直方图采样一致性(SAC-IA)算法基础上,通过加入同名点点对距离约束、几何形状约束,以及快速点特征直方图误差评定的改进,以此提高粗配准精度;精配准阶段,提出一种基于相邻站间大致重叠区域快速估算的最近点迭代(ICP)方法,该方法迭代收敛快、配准精度高。实验结果分析表明,该方法提高了配准精度和配准效率。
Terrestrial 3D laser point cloud registration is the foundation and key technology of point cloud analysis,three-dimensional reconstruction,and other engineering applications.In view of the shortcomings of the existing registration algorithm in terms of accuracy and efficiency,this paper proposes a registration method based on the fast point feature histogram and the calculation of the overlap area between two adjacent stations.In the coarse registration stage,based on the classical point feature histogram sample consensus initial alignment(SAC-IA)algorithm,by adding the same name point pair distance constraint,geometric shape constraint,and the improvement of the fast point feature histogram error evaluation,the coarse registration accuracy is improved.In the precise registration stage,a nearest point iteration(ICP)method based on the fast estimation of the approximate overlap area between adjacent stations is proposed,which has fast iterative convergence and high registration accuracy.The analysis of experimental results shows that the method in this paper improves the registration accuracy and efficiency.
作者
赵文峰
ZHAO Wenfeng(Shenzhen Changkan Investigation and Design Co.,Ltd.,Shenzhen 518003,China)
出处
《测绘工程》
2024年第4期68-75,共8页
Engineering of Surveying and Mapping
关键词
点云粗配准
点特征直方图
点云重叠区域
点云精配准
采样一致性
point cloud coarse registration
point feature histogram
point cloud overlapping region
point cloud precision registration
sample consensus initial alignment
