Identification of segment joint and automatic segmentation for shield tunnel based on LiDAR detection

Abstract

This study presents a novel method for identifying joints and automatically segmenting shield tunnels using light detection and ranging (LiDAR). In cylindrical coordinates, the Hough transform is used to extract feature LiDAR data corresponding to ring joints at different azimuths. This feature extraction using LiDAR data facilitates the computation of ring joint feature coordinates and average ring joint width. Subsequently, the M−estimator Sample Consensus (MSAC) algorithm is used to fit the plane containing the ring joint, resulting in successful recognition and segmentation of ring joints within the tunnel LiDAR data. Following the segmentation of the LiDAR data into distinct ring LiDAR data, the three-sigma (3σ) criterion is used to extract coordinates of longitudinal joint endpoints. The average width of the longitudinal joints is then determined. In cases where extraction of the longitudinal joint points is challenging, the azimuth difference in the design model is leveraged to calculate the azimuths. This approach enables joint recognition within LiDAR data as well as the geometric segmentation of individual segments. The proposed method is validated using a case study from Luoyang Metro Line 2. The results indicate that the segmentation method can accurately extract the majority of the ring and longitudinal joints. Moreover, these results are useful for not only monitoring structural health but also developing a building information model (BIM) for the tunnel.