论文部分内容阅读
针对激光扫描获取的古建筑点云数据在配准时存在的问题,提出了一种精确有效的多视图点云数据配准方法,使用NDT激光扫描匹配算法,对单个激光扫描重构2D离散数据点集进行相应的转换,将其转换为2维平面内分段连续可微的概率分布,同时选取Hessian矩阵法匹配其他的扫描,这样就能够有效解决点与点之间的对应问题。
In order to solve the existing problems in registration of ancient buildings’ point cloud data acquired by laser scanning, an accurate and effective multi-view point cloud data registration method is proposed. Using NDT laser scanning matching algorithm, a single laser scanning reconstructed 2D discrete data points Set the corresponding conversion, convert it into a continuously differentiable probability distribution of two-dimensional in-plane segments, and select the Hessian matrix method to match other scans, so that it can effectively solve the problem of the correspondence between points.