在利用机器人进行大尺寸曲面形貌测量的过程中,提出一种基于室内全球定位系统(iGPS)的点云拼接方法,以iGPS世界坐标系为点云拼接的坐标系,建立了点云拼接数学模型。利用粒子群优化(PSO)算法对迭代最近点(ICP)算法进行改进。基于球心距测量的点云拼接实验验证了所搭建测量系统的精度小于0.1 mm。在汽车前保险杠点云拼接实验中,最大负偏差为-0.05189 mm,最大正偏差为0.0727 mm,均小于0.1 mm,偏差分布较为均匀,验证了所提算法在大尺寸点云拼接方面具有较好的效果。

Motivated by the use of large-scale surface topography measurement by robots, we propose a method of point-cloud splicing based on the indoor global positioning system (iGPS). In our research, the iGPS world coordinate system is utilized as the coordinate system of point-cloud splicing to establish a mathematical model of point-cloud splicing. Furthermore, we employ the particle swarm optimization (PSO) algorithm for the iterative closest point (ICP) algorithm. The experimental results of point-cloud splicing of spherical distance measurement show that the accuracy of the measurement system is less than 0.1 mm. We also conduct a front-bumper point-cloud splicing experiment and the experimental result denote that the maximum negative deviation is -0.05189 mm and the maximum positive deviation is 0.0727 mm, which are less than 0.1 mm. It is also found that the deviation distribution is relatively uniform, which validates the proposed algorithm has a good effect on large-scale point-cloud splicing.