Automatic Detection Method for Shield Tunnel Segment Dislocation Based on Facility Point Cloud Removal and Segment...

Area of Science:

• Geotechnical Engineering
• Civil Engineering
• Geomatics Engineering

Background:

• Mobile laser scanning (MLS) is crucial for monitoring subway tunnel deformation.
• Segment dislocation is a key indicator of tunnel structural health, affecting mechanical and waterproof performance.
• Existing MLS methods face challenges with facility interference, seam localization, and automation.

Purpose of the Study:

• To develop an automated method for shield tunnel segment dislocation detection using MLS point cloud processing.
• To improve the accuracy and reliability of quantitative analysis for tunnel structural health monitoring.
• To address limitations of current MLS-based methods in facility interference and seam localization.

Main Methods:

• A point cloud filtering strategy using offset features and semantic segmentation to remove noise.
• A tunnel segment segmentation method combining bolt hole extraction and moving template matching for seam localization.
• Automated segment dislocation analysis based on filtering and segmentation results.

Main Results:

• The proposed filtering method shows significant accuracy improvements over conventional techniques.
• The method achieves less than 2 mm deviation in seam localization and dislocation analysis, showing high consistency with manual interpretation.
• The automated analysis is more accurate and reliable than existing approaches, reducing manual inspection needs.

Conclusions:

• The developed method enhances automation, consistency, and reliability in shield tunnel deformation assessment.
• It provides an effective solution for structural health monitoring of subway tunnels.
• The approach significantly reduces the reliance on manual inspection for quantitative analysis.