Monitoring tunnel deformations by means of multi-epoch dispersed 3D LiDAR point clouds: An improved approach

• High-accurate, automatic LiDAR technique was applied to tunnel deformation monitoring tasks.
• 3D MDP algorithm was proposed to estimate deformation signals for a tunnel surface.
• Improved results have been obtained compared to the existing 2D MDP algorithm.
• The quality of MDP signals was evaluated based on a rigorous covariance propagation approach.
• Comprehensive and reliable tunnel deformation monitoring has been successfully achieved.

Monitoring tunnel deformations is a crucial task when evaluating tunnel stability and safety. This task requires an accurate and high-resolution spatial technique to precisely capture the meticulous anomalies on a tunnel surface. As a response, the light detection and ranging (LiDAR) technique, which collects detailed spatial data in a fast and automatic manner, was recently proposed by Han et al. (2013) for monitoring the deformation of a 2D tunnel profile. Although the proposed approach successfully uses this modern spatial technique in tunnel analysis, the benefits of the 3D LiDAR technique have not been fully exposed. This study improved the technique as a real 3D approach. The associated uncertainties can be reduced by avoiding the 3D to 2D profile projection step. The minimum-distance projection (MDP) was then estimated using directly the 3D dispersed point clouds so that any deformation signal (point displacement) along the entire tunnel surface can be immediately identified. Furthermore, a rigorous covariance propagation approach was introduced to provide explicit quality indications on the obtained solution. The results of simulation tests and a real case study of a highway tunnel showed that the spatial implications of the 3D LiDAR technique can be fully explored by implementing the improved approach. Consequently, a more accurate and comprehensive solution for monitoring tunnel deformations can be achieved.