一种影像辅助的由粗到精的边坡点云自动配准

An image-aided coarse-to-fine registration for side slope point clouds

针对边坡激光扫描的多期点云自动化配准难度高、人工配准工作量大的问题,本文以提高边坡点云配准自动化程度及精度为出发点,提出一种由粗到精的边坡点云自动配准方法。首先利用激光虚拟影像完成两期激光点云粗配准;再根据多期地面激光扫描站心位置不变这一约束条件,提出带中心约束条件的ICP算法。实验表明,本文方法有效提高了边坡点云配准的效率与精度。

Automatic registration of side slope point clouds is difficult using existing methods, while manual registration is laborious. This paper proposed a coarse-to-fine registration to achieve accurate and automatic registration between multiple side slope point clouds. Firstly, coarse registration is fulfilled utilizing virtual laser scanning image. Secondly, a modified ICP algorithm is proposed according to the centers of the multiple scanning are identical. Experiments demonstrate the efficiency and accuracy of the proposed method.