Scholarlink

<p>Circumferential joint dislocation is a critical structural defect in subway shield tunnels, posing a serious threat to tunnel stability. Existing detection methods primarily rely on bolt-hole features, which are often blurred or indistinct in practice, thereby limiting their applicability. To address this challenge, we propose a bolt-hole-independent approach for accurate localization and dislocation detection of circumferential joints, based on point cloud intensity features and a block-based processing strategy. The proposed pipeline includes: (1) track separation via angular constraints; (2) removal of ancillary facilities using ellipse fitting and statistical filtering; (3) cylindrical projection; (4) intensity-based block segmentation; and (5) joint localization using Euclidean clustering with a two-step filtering process and translation correction of position. Dislocation detection is then performed based on the extracted joint positions. The proposed method was validated on two mobile laser scanning datasets acquired from operational subway shield tunnels—with and without distinguishable bolt-hole. Compared with existing approaches, our method demonstrated superior performance. Using manually measured results as ground truth, the dislocation detection mean accuracy reached 0.8&#xa0;mm and 0.9&#xa0;mm, respectively. The proposed method offers a direct and effective solution for circumferential joint dislocation without relying on bolt-hole.</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

A novel method for circumferential joint localization and dislocation detection in subway shield tunnels based on point cloud intensity features and block strategy

  • 期刊论文

  • 发表日期：

    2026年1月28日

查看全文
  • Shibiao Xu,
  • Qiuzhao Zhang,
  • Wei Li,
  • Wei Duan,
  • Kaikun Zhang,
  • Youquan Yang

摘要

Circumferential joint dislocation is a critical structural defect in subway shield tunnels, posing a serious threat to tunnel stability. Existing detection methods primarily rely on bolt-hole features, which are often blurred or indistinct in practice, thereby limiting their applicability. To address this challenge, we propose a bolt-hole-independent approach for accurate localization and dislocation detection of circumferential joints, based on point cloud intensity features and a block-based processing strategy. The proposed pipeline includes: (1) track separation via angular constraints; (2) removal of ancillary facilities using ellipse fitting and statistical filtering; (3) cylindrical projection; (4) intensity-based block segmentation; and (5) joint localization using Euclidean clustering with a two-step filtering process and translation correction of position. Dislocation detection is then performed based on the extracted joint positions. The proposed method was validated on two mobile laser scanning datasets acquired from operational subway shield tunnels—with and without distinguishable bolt-hole. Compared with existing approaches, our method demonstrated superior performance. Using manually measured results as ground truth, the dislocation detection mean accuracy reached 0.8 mm and 0.9 mm, respectively. The proposed method offers a direct and effective solution for circumferential joint dislocation without relying on bolt-hole.