摘要
针对迭代最近点(ICP)算法对初始位置敏感、收敛速度慢等问题,提出一种基于特征点与改进ICP的点云配准方法。新方法对原始点云进行体素栅格滤波,利用ISS算法分别提取源点云与目标点云的特征点,使用快速点特征直方图(FPFH)进行描述,并结合采样一致性算法(SAC-IA)求出初始变换,最终在粗配准基础上引入法向量约束改进ICP算法,完成点云精配准。实验以Bunny与Dragon点云作为测试对象,结果表明:改进算法可为精配准提供较为理想的初始位置,有效提高了配准的精度与速度。
Aiming at the problems that the iterative closest point algorithm is sensitive to the initial position and slow in convergence,a point cloud registration method based on feature points and improved ICP is proposed.In the method,the original point cloud is filtered by voxel grid,the feature points of the source point cloud and the target point cloud are extracted by ISS algorithm,described by fast point feature histogram,and combined with sampling consistency algorithm,the initial transformation is obtained.Finally,normal vector constraint is introduced to improve ICP algorithm on the basis of rough registration to complete the fine registration of point clouds.The experiment takes Bunny and Dragon point clouds as test objects,and the results show that the improved algorithm can provide ideal initial position for precise registration,and effectively improve the accuracy and speed of registration.
作者
黄际玮
陆安江
HUANG Jiwei;LU Anjiang(College of Big Data and Information Engineering,Guizhou University,Guiyang 550025,China)
出处
《微处理机》
2022年第6期38-42,共5页
Microprocessors
