Monitoring tunnel profile by means of multi-epoch dispersed 3-D LiDAR point clouds

Monitoring tunnel profiles is constantly one of the major routines when long-term tunnel stability is to be evaluated. Among many modern 3-D measurement techniques, the Light Detection and Ranging (LiDAR) technique is capable of collecting detailed spatial data in a fast and automatic manner and is thus a tool that can potentially be implemented for a tunnel monitoring task. However, raw LiDAR datasets usually contain non-specific and dispersive information and thus lack explicit geometric interpretation. In this study, an efficient approach for extracting tunnel profile geometry from raw LiDAR datasets will be presented. By utilizing in situ control features, tunnel profiles at multiple epochs can be automatically generated and expressed in a common datum definition. Then by applying a minimum-distance projection (MDP) algorithm, point correspondences are established so that deformation signals along any given profile can be immediately identified. Based on the results of simulation tests and a real case study of a highway tunnel, it is proven that the LiDAR technique makes possible an efficient and accurate solution for monitoring tunnel profile deformations when the proposed approach is implemented in such a task.

► High-accurate, automatic LiDAR technique was applied to tunnel profile monitoring.
► Algorithms were proposed to extract tunnel profiles using raw LiDAR datasets.
► Extracted profiles provide accurate and detailed representations of tunnel sections.
► Millimeter-level deformation signals of monitored profiles have been successfully detected.
► Required field work and traffic disturbance can be significantly reduced by the proposed.