三维点云ICP算法改进研究
周春艳;李勇;邹峥嵘
【期刊名称】《计算机技术与发展》 【年(卷),期】2011(021)008
【摘要】Rapid development of three-dimensional laser scanning technology is to make it wider application in all fields. Due to physical limits, a three-dimensional laser scanning can't acquire all the data object, so to scanning point clouds for stitching. First of all, conduct a series of studies on ICP algorithm most frequently used in this document, a prerequisite for ICP algorithm is a good registration of initial value, the article on registration of initial value selection uses principal component analysis method, as the follow-up work to provide a good prerequisite of ICP algorithm to increase point of pretreatment and to seek to increase various restrictions, kd-tree acceleratedly finds to improve algorithm, and through an instance to verify the validity of this algorithm and rationality.%三维激光扫描技术的快速发展,使其在各个领域得到广泛应用.由于物理上的一些限制,一次三维激光扫描不能获取扫描物体的全部数据,因此要对扫描点云进行拼接.首先,对最常用的ICP算法进行一系列研究,ICP算法的前提条件是具有一个良好的配准初值,文中在配准初值的选取上采用主成分分析法,为后续ICP算法的工作提供一个良好前提条件,增加点集预处理,点对查找上增加各种限制,采用kd-tree加速查找,以此对算法进行改进,并通过实例来验证本算法的有效性及合理性.
三维点云ICP算法改进研究
