基于FPFH特征的ICP点云配准的改进算法  

Improved ICP Point Cloud Registration Algorithm Based on FPFH Features

作　　者：宗晓萍[1] 吴岩 

机构地区：[1]河北大学电子信息工程学院,河北保定071000

出　　处：《工业控制计算机》2021年第7期106-108,共3页Industrial Control Computer

基　　金：河北省科技计划项目:基于物联网的冷链物流协调控制及优化(15214712)。

摘　　要：针对ICP配准技术速度慢、对初始位置敏感的问题,提出一种基于FPFH特征的ICP点云配准改进算法。在提取点云的FPFH作为点云特征描述、使用SAC-IA作为初始配准的基础上,利用均匀降采样对输入点云进行数据精简,在寻找最近点过程中应用Kdtree数据结构加快寻找速度,应用线性最小二乘法优化ICP配准,最终得到从源点云到目标点云的变换矩阵。实验结果表明,改进的ICP算法可以在提高配准精度的同时加快配准速度。Aiming at the problem of slow ICP registration technology and sensitivity to initial position,an improved ICP point cloud registration algorithm based on FPFH features is proposed in this paper.On the basis of extracting the FPFH of the point cloud as the point cloud feature description and using SAC-IA as the initial registration,the input point cloud is reduced by uniform downsampling,and the Kdtree data structure is used to speed up the search in the process of finding the closest point.The linear least squares method is applied to optimize the ICP registration,and finally the transformation matrix from the source point cloud to the target point cloud is obtained.

关 键 词：点云配准 FPFH特征描述 ICP算法 改进算法 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]