关键词: 形变检测, 点云数据, 轴线拟合, 隧道分割

Abstract:

In order to solve tunnel deformation detection, the spatial axis extraction, tunnel intensity image generation and segmentation, tunnel point cloud segmentation and tunnel section deformation detection were carried out. For the raw point cloud of the tunnel obtained by the 3D laser scanner, an improved RANSAC algorithm was first proposed to fit the center of the equal-distance tunnel sections, based on the designed axis, and then a fitting method of cubic B-spline curve was applied to obtain the spatial axis of the tunnel. Further, the tunnel intensity image was generated by applying an algorithm of data transformation, and the boundary detection and geometric weights calculation algorithms were performed to extract all boundary lines in the image. By mapping the boundary line pixels to the three-dimensional space, the segmentation of the tunnel point cloud was achieved. Finally, an accurate detection metrics for tunnel section deformation was proposed. The results show that this algorithm is able to segment the tunnel point cloud accurately and improve the accuracy and reliability of tunnel section deformation detection, which has the high practical value.

Key words: deformation detection, point cloud data, spatial axis extraction, tunnel segmentation