Open Access
ARTICLE
Vision-Based Intelligent Detection of Nonuniform Tunnel Excavation Contours and Stress Response Characteristics of Support Structures
Liuyang Sun1, Jinfang Xue1, Wu Zhou1, Yujiang Lü2,*, Junsheng Yang2, Cong Zhang3, Xiong Cao3
1 Civil Engineering Company, Sinohydro Bureau 14 Co., Ltd., Kunming, China
2 School of Civil Engineering, Central South University, Changsha, China
3 College of Soil and Water Conservation, Central South University of Forestry and Technology, Changsha, China
* Corresponding Author: Yujiang Lü. Email: 
(This article belongs to the Special Issue: Damage Recognition and Health Monitoring for Transportation Infrastructure)
Structural Durability & Health Monitoring https://doi.org/10.32604/sdhm.2026.089636
Received 23 July 2026; Accepted 08 September 2026; Published online 23 September 2026
Abstract
This study investigates the vision-based intelligent detection of nonuniform tunnel excavation contours and the stress response of support structures under mechanized drill-and-blast construction, using a high-altitude tunnel project as a case study. The nonuniform contour is primarily caused by the outward inclination of perimeter holes imposed by the operating envelope of a drilling jumbo. Field image acquisition, sparse point-cloud reconstruction, dense point-cloud generation, and three-dimensional surface reconstruction were integrated to reconstruct the actual tunnel wall after blasting and identify its geometric characteristics. The measured outward inclination angles of the perimeter holes were generally stable at approximately 4.5°, consistent with the specified control range. Numerical models based on the measured nonuniform contour and the conventional uniform contour were then established to compare the stress and displacement responses of the primary support. The two excavation conditions exhibited similar response patterns, and both satisfied the strength and deformation-control requirements. Field monitoring showed that the maximum surrounding-rock displacement, rock-bolt stress, and shotcrete stress were approximately 2.6 mm, 13 and 2.0 MPa, respectively, and all monitored responses evolved smoothly toward stabilization. The results demonstrate that the proposed vision-based method can effectively identify nonuniform excavation contours. Under the investigated Grade III rock-mass conditions, with an outward inclination angle of approximately 4.5° and the support parameters adopted in this study, the mechanical response of the primary support remained controllable, providing a practical reference for excavation-quality assessment and support design in similar tunnel projects.
Keywords
Nonuniform excavation; outward inclination angle; vision-based intelligent detection; three-dimensional reconstruction; support response
Login
Register
Submit a Paper
Propose a Special lssue
Download PDF
Downloads
Citation Tools
