基于PCA融合优化的GICP点云配准算法

GICP Point Cloud Registration Algorithm Based on PCA Fusion Optimization

文章提出一种广义迭代最近点(GICP)算法,对三维点云数据进行背景去除、剔除噪声点及体素滤波降采样预处理,计算点云协方差矩阵得到主特征分量,找到两组点云的主轴方向并进行校正,完成点云的初始配准,利用K–D tree搜索临近点并根据初始配准提供的位姿参数对点云进行GICP精配准。实验表明,该算法在配准精度、算法收敛速度及迭代次数上均优于传统ICP算法。

This paper proposes a generalized iterative nearest point(GICP)algorithm,which performs background removal,noise removal and voxel filtering downsampling pre-processing on 3D point cloud data,calculates the point cloud covariance matrix to obtain the main feature component,finds the principal axis direction of two groups of point clouds and makes correction,and completes the initial registration of point clouds.The K-D tree was used to search the adjacent points and perform GICP fine registration of the point cloud according to the pose parameters provided by the initial registration.Experiments show that the algorithm is superior to the traditional ICP algorithm in registration accuracy,convergence speed and iteration times.