Scholarlink

<p>Tunnel infrastructure management increasingly demands efficient and accurate geometric modeling techniques, which serve as a critical foundation for Structural Health Monitoring (SHM) tasks. Laser Scanning System (LSS) shows its potential to provide wide-range, stable 3D sensing for SHM. However, it is challenging for 3D reconstruction in tunnels due to satellite signal dropouts, degraded perception, and observation degeneracy. Consequently, this research proposes a multi-sensor fusion SLAM technique to achieve accurate modeling of the tunnel environment. In detail, we developed a LiDAR-inertial system (Robust-LIWO) that integrates information from LiDAR, IMU, and wheel odometry using an Error State Kalman Filter (ESKF). In the system, a degeneracy detection method based on point cloud constraint relations is utilized for degeneracy analysis and a dual-LiDAR fusion strategy for high-quality 3D reconstruction. A field experiment is carried out in an actual tunnel. Results show that Robust-LIWO achieves high-precision localization and 3D reconstruction with maximum error of 0.16m and trajectory error RMSE of 0.10m, respectively. The Robust-LIWO provides an efficient and accurate modeling tool for the digital twins of tunnel engineering.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Efficient and accurate 3D reconstruction in a featureless tunnel environment: a robust LiDAR-Inertial SLAM tightly coupled with wheel odometry

  • 期刊论文

  • 发表日期：

    2025年8月30日

查看全文
  • Wenli Liu,
  • Yafei Qi,
  • Tanghao Guan,
  • Yaofeng Liang

摘要

Tunnel infrastructure management increasingly demands efficient and accurate geometric modeling techniques, which serve as a critical foundation for Structural Health Monitoring (SHM) tasks. Laser Scanning System (LSS) shows its potential to provide wide-range, stable 3D sensing for SHM. However, it is challenging for 3D reconstruction in tunnels due to satellite signal dropouts, degraded perception, and observation degeneracy. Consequently, this research proposes a multi-sensor fusion SLAM technique to achieve accurate modeling of the tunnel environment. In detail, we developed a LiDAR-inertial system (Robust-LIWO) that integrates information from LiDAR, IMU, and wheel odometry using an Error State Kalman Filter (ESKF). In the system, a degeneracy detection method based on point cloud constraint relations is utilized for degeneracy analysis and a dual-LiDAR fusion strategy for high-quality 3D reconstruction. A field experiment is carried out in an actual tunnel. Results show that Robust-LIWO achieves high-precision localization and 3D reconstruction with maximum error of 0.16m and trajectory error RMSE of 0.10m, respectively. The Robust-LIWO provides an efficient and accurate modeling tool for the digital twins of tunnel engineering.