Tunnel Leakage Detection Method based on Improved YOLOv8
DOI:
https://doi.org/10.6919/ICJE.202506_11(6).0034Keywords:
Tunnel Leakage Detection; YOLOv8; Dynamic Snake Convolution; BiFPN.Abstract
Tunnel leakage poses significant threats to structural integrity and operational safety. Traditional detection methods, such as manual inspections and instrument-based techniques, are inefficient and costly, particularly in complex tunnel environments with irregular leakage morphologies, varying illumination, and background interference. This paper proposes YOLOv8-LD, an enhanced version of the YOLOv8-seg model, specifically designed for tunnel leakage detection. By integrating DySnakeConv for adaptive edge feature extraction, SegNext attention mechanism for robust feature fusion, and BiFPN for multi-scale feature integration, YOLOv8-LD addresses these challenges effectively. A custom dataset of 3,000 annotated tunnel leakage images was developed to train and evaluate the model. Experimental results demonstrate that YOLOv8-LD achieves a mean Average Precision (mAP@50) of 73.7%, a 9.2% improvement over the baseline YOLOv8-seg, while maintaining real-time performance at 104 FPS. Ablation studies and visualization analyses confirm the model's superior accuracy, robustness, and suitability for practical tunnel maintenance.
Downloads
References
[1] Kaartinen E, Dunphy K, Sadhu A. LiDAR-based structural health monitoring: Applications in civil infrastructure systems[J]. Sensors, 2022, 22(12): 4610.
[2] Wang J, Zhang J, Cohn A G, et al. Arbitrarily-oriented tunnel lining defects detection from Ground Penetrating Radar images using deep Convolutional Neural networks[J]. Automation in Construction, 2022, 133: 104044.
[3] Xu Y, Li D, Xie Q, et al. Automatic defect detection and segmentation of tunnel surface using modified Mask R-CNN[J]. Measurement, 2021, 178: 109316.
[4] Hussain M. YOLO-v1 to YOLO-v8, the rise of YOLO and its complementary nature toward digital manufacturing and industrial defect detection[J]. Machines, 2023, 11(7): 677.
[5] Feng S J, Feng Y, Zhang X L, et al. Deep learning with visual explanations for leakage defect segmentation of metro shield tunnel[J]. Tunnelling and Underground Space Technology, 2023, 136: 105107.
[6] Wu Y, Han Q, Jin Q, et al. LCA-YOLOv8-Seg: an improved lightweight YOLOv8-Seg for real-time pixel-level crack detection of dams and bridges[J]. Applied Sciences, 2023, 13(19): 10583.
[7] Yang K, Bao Y, Li J, et al. Deep learning-based YOLO for crack segmentation and measurement in metro tunnels[J]. Automation in Construction, 2024, 168: 105818.
[8] Guo M H, Lu C Z, Hou Q, et al. Segnext: Rethinking convolutional attention design for semantic segmentation[J]. Advances in neural information processing systems, 2022, 35: 1140-1156.
[9] Doherty J, Gardiner B, Kerr E, et al. BiFPN-yolo: One-stage object detection integrating Bi-directional feature pyramid networks[J]. Pattern Recognition, 2025, 160: 111209.
[10] Behbahani S S, Golrokh A J, Hafiz A, et al. Utilizing infrared thermography for the condition assessment of tunnel Lining with tiled surface in various temperature conditions[J]. Tunnelling and Underground Space Technology, 2024, 154: 106093.
[11] Wang P, Wang S, Jierula A. Automatic identification and location of tunnel lining cracks[J]. Advances in Civil Engineering, 2021, 2021(1): 8846442.
[12] Nhat-Duc H, Nguyen Q L, Tran V D. Automatic recognition of asphalt pavement cracks using metaheuristic optimized edge detection algorithms and convolution neural network[J]. Automation in Construction, 2018, 94: 203-213.
[13] Man K, Liu R, Liu X, et al. Water leakage and crack identification in tunnels based on transfer-learning and convolutional neural networks[J]. Water, 2022, 14(9): 1462.
Downloads
Published
Issue
Section
License
Copyright (c) 2025 International Core Journal of Engineering
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.
How to Cite
