Defects detection of shield tunnel segment based on terrestrial laser point cloud

The defects of shield tunnel segments, such as dislocation, cracking and falling off, are one of the hot research topics in underground engineering. Based on data driven thoughts, the point cloud of shield tunnel is expanded into a binary image. Due to the image contains the mapping topological relationship between the original point cloud and the pixel, the morphological algorithm was used to identify the splice joints between rings, segmentation of tunneling image expansion is realized ring by ring. Then, the laser point cloud data were segmented at the segment scale, and the defects could be identified and detected in the segment. The experimental results show that the proposed method could effectively identify the segment defects.