点云配准是三维激光扫描数据处理过程中不可或缺的一个环节,利用标靶进行配准是经典的手段之一,此类方案在单独扫描标靶的基础上进行半自动化配准。本文给出一种自动配准策略,用中心投影原理将单站扫描的点云转换为深度影像,借助数字图像处理技术完成标靶的自动提取,拟合获得标靶中心点的坐标,并借用摄影测量学的知识实现点云的自动化配准。实验证明了本文方法的可靠性。 Point cloud registration is an indispensable link in 3D laser scanning data processing. Registration using the target is one of the classic methods. Semi-automatic registration is based on the single scanning target. In this paper, an automatic registration strategy is proposed, which converts the point cloud of a single station to a depth image using the principle of central projection, completes the automatic extraction of the target by using digital image processing technology, obtains the coordinates of the target center point by fitting and borrows Photogrammetry knowledge to automate point cloud registration. The experiment proves the reliability of this method.