摘要
本文提出了一种基于曲率特征点初配准的多源点云融合解决方案。其次使用迭代最近点算法(ICP)采取点到面的方法对点云数据进行精配准,在K维树(K-d树)的最近值领域搜索的基础上使用点云配准的3倍中误差作为判断条件,交叉验证保多源点云数据中同名点云数据。小于判断条件时保留三维激光点云数据的空间位置并采用加权平均思想过渡点云颜色。本文以倾斜摄影测量中易拉花的复杂独立地物“鲁班像”为研究对象,使用上述方法解决了多源点云数据融合时数据冗余与色差明显的问题,融合后“鲁班像”模型颜色自然、线条顺滑、棱角明显。
This article proposes a multi-source point cloud fusion solution based on initial registration using curvature feature points.The iterative Closest Point(ICP)algorithm is used to achieve precise point cloud registration through a point-to-plane method.Based on K-dimensional tree(K-d tree)nearest neighbor search,the 3-time registration error of point cloud is used as the judging condition to cross-validate and retain the overlapping data in multi-source point cloud data.When the error is below the specified threshold,the spatial coordinates of three-dimensional(3D)light detection and ranging(LiDAR)point cloud data are retained and a weighted average method is used to transition point cloud color.The research subject is the complex independent object“Luban statue”in oblique photography measurement.The proposed approach effectively solves the problems of data redundancy and color aberrations in multi-source point cloud fusion,resulting in a fused“Luban statue”model with natural color,smooth lines,and clearly defined edges.
作者
李闰
陈胜雷
闫全超
LI Run;CHEN Shenglei;YAN Quanchao(School of Municipal Engineering,Hubei Urban Construction Vocational and Technological College,Wuhan Hubei 430000,China;Heze Urban Constructing Group,Heze Shandong 274000,China;Shaanxi Bureau of Surveying,Mapping and Geoinformation,Xi'an Shaanxi 710000,China)
出处
《北京测绘》
2023年第11期1486-1490,共5页
Beijing Surveying and Mapping
关键词
倾斜摄影测量
三维激光扫描
多源数据
点云融合
曲率特征点
oblique photogrammetry
three-dimensional(3D)laser scanning
multi-source data
point cloud fusion
curvature eatures point
