Centerline-based registration for shield tunnel 3D reconstruction using spinning mid-range LiDAR point cloud and multi-cameras

Abstract

Mobile measurements can rapidly acquire tunnel information. However, cumulative errors in yaw angles occur in the absence or weakness of global positioning system (GPS) signals. This paper presents a method for 3D reconstruction of shield tunnels based on tunnel centerlines using non-repeating spinning mid-range LiDAR (SML) points and photos. First, a low-cost mobile measurement system (MMS) was built. Subsequently, the raw data were transformed into the tunnel centerline coordinate system (TCCS), including coarse registration with centerline alignment and fine registration based on convex hull areas. The multi-sensor data were fused in the TCCS, and photos were projected onto the SML points and unwrapped. Kilometrage corrections were applied by weighting the errors between the survey control points on panoramic images and their geodetic coordinates. Finally, the reconstructed data were located using image segmentation and indexing. This approach demonstrates higher registration accuracy in subway scenes than mainstream algorithms.