Localization in Tunnels Using Feature-based Scan Matching

In tunnel inspections, it is important to determine whether cracks and other deformations are caused by external forces. Therefore, point clouds at two different periods are compared to determine if any deformation has occurred and register these point clouds. In mobile mapping systems using global navigation satellite system (GNSS)/inertial measurement unit (IMU) navigation, tunnel point clouds are distorted due to accumulated errors. Therefore, it is not possible to compare the acquired raw point clouds with each other. Previously, point clouds were aligned manually, but this was time-consuming and expensive. In this work, we propose a positioning correction technique based on scan matching according to tunnel features and describe the evaluation results. The evaluation experiment showed the proposed method’s effectiveness, reducing the average error from 0.834 m to 0.063 m.