The three dimensional deformation and corresponding temporal evolution of surrounding rock is a key information for tunnel stability analysis and safety warning. To reconstruct the 3D deformation process of tunnel based on point cloud data by laser scanning, a semi-automatic method is presented to align the point cloud data of the tunnel surfaces at different stages. This method includes three steps: (i) initial alignment, splitting, denosing and merging of the tunnel surface’s point clouds; (ii) coordinate transformation between the tunnel surface’s point clouds according to the arch’s geometric characteristics; and (iii) calculation and visualization of the tunnel surface’s sequential deformation based on the multiscale model cloud comparison algorithm. The technical application of the time-dependent deformation analysis for an experimental tunnel in a mine indicated that this method can not only fully expose the evolution process of tunnel surrounding rock deformation, i.e., the spatial inhomogeneity, partitioning and temporal characteristics, but also can warn instability of tunnel by discerning hot deformation islands, i.e., zones with relatively large deformation.

围岩的三维变形及相应的时间演化是隧道稳定性分析和安全预警的关键信息。为了通过激光扫描基于点云数据重建隧道的3D变形过程，提出了一种半自动方法，对隧道表面点云数据在不同阶段进行对准。该方法包括三个步骤：（i）隧道表面的点云的初始对准，分割，去噪和合并；（ii）根据拱的几何特征，在隧道表面的点云之间进行坐标转换；（iii）基于多尺度模型云比较算法计算和可视化隧道表面的顺序变形。