一种精简点云的快速配准算法

A quick registration algorithm based on reduced point cloud

针对三维重建过程中点云数据存在的配准速度慢、精度低问题,提出一种精简点云的快速配准算法。先采用自适应栅格法对点云数据进行精简;在点云数据精简的基础上,运用快速ICP算法进行配准,即采用法向量的变化提取初始特征点,并使用RANSAC算法获取初始配准点集,完成粗匹配;再通过最小二乘法迭代计算最优的坐标变换,完成点云的精匹配。实验结果表明,所提算法在保证良好的原始点云数据的几何特征的同时,有效减少点云数据的冗余量,提高了运行效率和匹配精度,为实时三维重建提供了有效保证。

Aiming at the problem of slow matching speed and low accuracy of point cloud data in the process of 3 D reconstruction,a fast point cloud registration algorithm based on voxel grid method is proposed. Firstly,the adaptive voxel grid method is used to simplify the point cloud data. Based on this,the quick ICP algorithm is used to register.In order to complete the rough matching,the normal vector is used to extract the initial feature points,and RANSAC algorithm is used to obtain the initial registration point set. Then the optimal coordinate transformation is iteratively calculated by the Least Squares algorithm for completing the fine matching of the point cloud. The experimental results show that the proposed algorithm can effectively reduce the redundancy of point cloud data while ensuring good geometric characteristics of the original point cloud data,improve the operating efficiency and matching accuracy,and provide an effective guarantee for real-time three-dimensional reconstruction.