地面三维激光扫描点云的多站数据无缝拼接

Seamless Multi Station Merging of Terrestrial Laser Scanned 3D Point Clouds

针对如何将多站分块扫描的点云无缝地拼接为一个整体以完整地表达扫描对象几何形态的问题,基于测量平差理论,提出了一种简单易行的多站扫描数据无缝拼接方法.首先根据与相邻前一扫描站相重叠的扫描区域内的同名标志点按附有限制条件的间接平差模型求取坐标转换参数,再对序列坐标转换参数按闭合条件进行加权误差分配,得到各站扫描点云的最终坐标转换模型,并依此坐标转换模型对各站扫描点云进行坐标变换,将它们统一到项目坐标系下,最终得到多站扫描点云的无缝拼接结果.对多站扫描的建筑物点云数据进行整体拼接实验,证明了本方法的可行性和有效性.

Aiming at solving the problem of how to merge the 3D point clouds acquired by terrestrial laser scanner at different stations to form the whole representation of scanned object, a simple and seamless data merging approach for data from multi stations is proposed based on the theory of surveying adjustment. Firstly, the parameters of coordinate transformation are worked out with restricted conditional parameter adjustment, which uses the homologous points in overlapped region of this station and the previous neighbor station as the thread. Secondly, the parameters of each station are further adjusted serially and weightedly considering the closure condition. The final coordinate transformation models for each station can be set up with these adjusted parameters, and the coordinate of point clouds acquired at each station are transformed to the same project coordinate system and merged to a whole. The seamless merging result which is carried out on a building scanned at multi stations indicates that this data merging approach is feasible and effective.