Synapse
⌘+K
Synapse
PulseExploreClubsResearchersJournals
Instagram
HomeClubsExplore
March 29, 2026Automation in ConstructionOpen Access

Tunnel scanner: Geometry-informed synthetic point cloud generation and transfer learning for tunnel segmentation

View Full Paper
Ask AI
Bookmark
Share

Authors

WYWanru YangUniversity of CambridgeWLWei LinYJYixiong JingUniversity of Cambridge

Discussion

Loading...

Member takes

Implication

This research develops a simulator that improves tunnel segmentation accuracy using synthetic data and machine learning techniques.

Key Points

  • The aim is to enhance tunnel segmentation using synthetic point clouds and deep learning methods.
  • Developed a high-fidelity simulator for generating synthetic tunnel point clouds.
  • Implemented a hybrid position-normal local spatial encoding module for geometric feature integration.
  • Utilized transfer learning to improve performance on tunnel segmentation tasks.
  • Models trained solely on synthetic data achieved 71.9% mean Intersection-over-Union (mIoU) and 86.6% overall accuracy (OA).
  • Transfer learning increased performance to at least 78.8% mIoU and 90.9% OA with limited real data.
  • The approach resulted in a 24% increase in segmentation accuracy with constrained real datasets.

Cite This Study

Yang et al. (2026) studied this question.

synapsesocial.com/papers/69c8c399de0f0f753b39e7c8https://doi.org/10.1016/j.autcon.2026.106924
View Full Paper
Ask AI
Bookmark
Share

Also Consider

Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context:

  1. 1Semantic segmentation of large-scale segmental lining point clouds using 3D deep learning2024 · 2 citations
  2. 2Attention-Guided Semantic Segmentation and Scan-to-Model Geometric Reconstruction of Underground Tunnels from Mobile Laser Scanning2026 · 1 citations
  3. 3Deep Learning–Based Few-Shot Method for Detection and Evaluation of Tunnel Overbreak and Underbreak2026
  4. 4Denoising image point clouds using segmentation and synthetic data for enhanced structural health analysis of tunnels2026
  5. 53D Scene Understanding for Highway Tunnel Construction: Challenges and Baseline Analysis2025