A Roughness, Feature-based Algorithm of Point Cloud Segmentation of the Secondary Lining in NATM Tunnel

Abstract

To realize point cloud segmentation of secondary lining in tunnel under construction, a roughness feature-based point cloud segmentation algorithm for secondary lining is developed in this paper. Preprocessing of raw scanned data enables the parameterization of the optimal voxel mesh size, roughness calculation radius and extraction threshold with algorithm experiments. Primary segmentation is performed by clustering under optimal parameters, and then the results of the primary segmentation are input as seed points for secondary segmentation. The effective segmentation rate is up to 99.25%. Comparison confirms that the performance of the proposed algorithm is superior to other similar algorithms, proving its’ practicality and feasibility.