古遗址多视点云的渐进式三维配准

Sequence iterative 3D registration of multi-view point cloud for archaeological site

为实现古遗址场景多分辨率点云的精确配准,提出一种基于序列迭代的多视点云三维配准方法。利用反向投影确定两点集的重叠区,基于法向量计算定标球球心坐标,实现相邻点云的准确融合;利用基于马氏距离计算相邻点云的三维边缘点,通过在迭代过程中不断优化转换矩阵,逐步提高配准精度,消除部分重影现象。唐小雁塔和西汉古墓的工程实践表明,该方法能有效抑制序列配准的累积误差,具有有效性和鲁棒性,可应用于古遗址的数字化保护与虚拟展示领域。

Aiming at the multi-resolution point cloud registration problems of the archaeological site scene,this paper put forward a 3D cloud registration method based on sequence iterative.Firstly,based on the synchronous scan image,a rough registration of the adjacent point cloud could be obtained in an interactive way.Secondly,based on a normal-based algorithm of registration sphere center,filtered the points of sphere surface from all scanned point cloud data according to the normal of data points and characteristic of viewpoint,then the center of registration sphere could be calculated accurately,the initial registration of the adjacent point cloud could be obtained.Finally,based on the 3D edge points of the adjacent point cloud from Markova distance calculations,some overlapping images could be eliminated by the way of constantly optimizing the transition matrix in the iteration process.The engineering practice of the Small Wild Goose Pagoda in Tang Dynasty and the Ancient Tomb in Han Dynasty prove that the method can effectively restrain the accumulative errors of sequence registrations and can be applied to the archaeological site of digital protection and virtual exhibition fields.