一种新的扫描点云自动配准方法

点云数据配准是三维激光数据处理的关键问题,结合曲率进行配准是点云配准非常有效的一种方式。为了提高配准的速度和精度,针对点云曲率和RANSAC算法的特点和应用,分析了两种方法之间的优缺点。研究了一种结合曲率的RANSAC点云配准算法,并利用改进的ICP算法对点云进行精确配准。结果表明:基于曲率的RANSAC算法能够显著提高初始配准的速度和精度,同时改进的ICP算法比原有ICP算法提高了二次配准的速度和精度。

A New Automatic Registration Method of Scanning Point Cloud

The point cloud data registration is the key issue of 3-d laser data processing.The point cloud registration combined with curvature distribution is a very effective away.In order to improve the registration speed and precision,in view of the characters and application of point cloud curvature and RANSAC algorithm,the advantages and disadvantages of the two methods are analyzed here.Then,a RANSAC point cloud registration algorithm combined with curvature is studied,and the improved ICP algorithm is used for point cloud registration.The results show that: the RANSAC algorithm based on curvature can significantly improve the initial registration speed and precision,at the same time,the improved ICP algorithm can significanlly improve the secondary registration speed and precision compared with the original ICP algorithm.