一种基于降采样后关键点优化的点云配准方法

Point cloud registration method based on key point optimization after downsampling

针对工件点云数据多而导致点云配准耗时长的问题,提出一种基于降采样后关键点优化的点云配准方法。计算点云若干体素的重心,利用kd-tree快速遍历重心的邻近点来代替该体素;提出自适应的点云平均距离计算方法,对降采样后的点云提取ISS3D关键点,并采用基于球邻域的边界点判断方法对其优化;对优化后的关键点进行FPFH特征描述,利用SAC-IA求解近似变换阵,使用ICP算法精配准而解得工件的精确位姿信息。实验结果表明,相较于其他四种配准算法,配准精度分别提高了96.9%、98.1%、93.3%和3.5%,配准速度分别提高了77.2%、77.7%、76.9%和85.4%,表明了该方法的有效性。

Aiming at the problem of time-consuming point cloud registration due to the large amount of point cloud data of the workpiece,this paper proposed a point cloud registration method based on key points optimization after downsampling.It calculated the center of gravity of several voxels in the point cloud and used kd-tree to quickly traverse the neighboring points of the center of gravity to replace the voxel.It proposed an adaptive point cloud average distance calculation method to extract the ISS3D key points from the down-sampled point cloud,and used boundary point judgment method of the spherical neighborhood to optimize.The characteristic of FPFH described the key points after optimized.It used SAC-IA to solve approximate transformation matrix,and used ICP algorithm for precise registration to obtain the precise pose information of the workpiece.The experimental results show that compared with the other four registration algorithms,the registration accuracy has improved by 96.9%,98.1%,93.3%,and 3.5%,the registration speed has increased by 77.2%,77.7%,76.9%,and 85.4%,it shows the effectiveness of the method.