Synapse
⌘+K
Synapse
PulseExploreClubsResearchersJournals
Instagram
HomeClubsExplore
April 23, 2026Computer-Aided Civil and Infrastructure EngineeringOpen Access

Maintenance-oriented tunnel digital model generation via panoptic segmentation of ultra-high-resolution images

View Full Paper
Ask AI
Bookmark
Share

Authors

ZYZehao YeQLQi LiGDGiuseppe Desiderio

Discussion

Loading...

Member takes

Overview

Novel framework automates damage detection in ultra-high-resolution tunnel images, enhancing inspection reporting.

Key Points

  • The aim is to develop an automated framework for damage detection and 3D reconstruction using ultra-high-resolution tunnel images.
  • Utilized ultra-high-resolution panoramic images for direct damage detection and 3D reconstruction.
  • Employed panoptic segmentation with a flexible architecture for detailed context extraction.
  • Integrated Building Information Modelling for creating an annotated digital model of tunnel defects.
  • Increased panoptic quality from 53.14 to 56.99 using the Segment Anything Model 2 backbone.
  • Demonstrated efficient detection of five main damage types within tunnel components.
  • Supported creation of an as-damaged BIM model for better monitoring and management of tunnels.

Cite This Study

Ye et al. (2026) studied this question.

synapsesocial.com/papers/69e9baa885696592c86ecbe9https://doi.org/10.1016/j.cacaie.2026.100032
View Full Paper
Ask AI
Bookmark
Share