多尺度边缘信息融合的隧道渗漏水检测方法

doi:10.11947/j.AGCS.2026.20260007

摘要/Abstract

摘要：

随着城市轨道交通的快速发展，地铁隧道渗漏水检测需求日益增长，隧道衬砌表面存在较多附属物，对渗漏水区域造成遮挡，且部分渗漏水边界模糊，给检测带来较大挑战。现有方法在复杂隧道背景下对渗漏水边缘特征捕捉能力不足，导致边界模糊或形态多变的渗漏水区域检测精度和稳健性受限制。针对上述问题，本文提出一种基于多尺度边缘信息融合的隧道渗漏水检测方法。该方法由多特征融合图像增强算法和GLEF-YOLO检测模型组成。在图像预处理阶段，融合小波变换增强的各向异性扩散滤波与自适应直方图均衡化，增强渗漏水区域的边缘细节和局部对比度；在检测模型中，引入多尺度边缘信息传递模块（MEIT）和全局-局部感知融合模块（GLAFusion），分别用于跨层传递浅层边缘信息，并融合局部细节与全局语义特征，从而提升模型对复杂背景下不规则渗漏水目标的识别能力。试验结果表明，GLEF-YOLO比基线模型的mAP提升了8.4个百分点。与近期优秀的目标检测方法进行比较，本文方法在各项指标上都达到了最优效果，并且在复杂背景的渗漏水检测中精度更高，对管线、附属物遮挡及背景纹理干扰的稳健性更强，验证了本文方法的有效性，为隧道渗漏水的智能化、高精度检测提供了有效的解决方案。

关键词: 隧道渗漏水, 图像增强, 目标检测, 三维激光扫描, 多尺度边缘信息传递

Abstract:

With the rapid development of urban rail transit, the demand for water leakage detection of subway tunnels is increasing. There are many attachments on the surface of tunnel lining, which block the water leakage area, and some of the water leakage boundaries are blurred, which brings great challenges to the detection. The existing methods have insufficient ability to capture the edge features of water leakage in complex tunnel background, which limits the detection accuracy and robustness of water leakage areas with blurred boundaries or variable shapes. In view of the above problems, this paper proposes a tunnel water leakage detection method based on multi-scale edge information fusion. The method consists of multi-feature fusion image enhancement algorithm and global-local edge information fusion YOLO (GLEF-YOLO) detection model. In the image preprocessing stage, the wavelet transform enhanced anisotropic diffusion filtering and adaptive histogram equalization are combined to enhance the edge details and local contrast of the water leakage area. In the detection model, a multi-scale edge information transfer module (MEIT) and a global-local perception fusion module (GLAFusion) are introduced to transmit shallow edge information across layers, and local details and global semantic features are fused to improve the recognition ability of the model for irregular water leakage targets in complex backgrounds. The experimental results show that the mAP of GLEF-YOLO is 8.4 percentage points higher than that of the baseline model. Compared with the recent excellent target detection methods, it has achieved the best results in all indicators, and has higher accuracy in the detection of water leakage in complex backgrounds, and is more robust to pipeline, appendage occlusion and background texture interference. The effectiveness of the proposed method is verified, which provides an effective solution for the intelligent and high-precision detection of tunnel water leakage.

Key words: tunnel water leakage, image enhancement, target detection, 3D laser scanning, multi-scale edge information transfer

中图分类号:

P237

刘硕, 孙海丽, 钟若飞. 多尺度边缘信息融合的隧道渗漏水检测方法[J]. 测绘学报, 2026, 55(6): 1116-1127.

Shuo LIU, Haili SUN, Ruofei ZHONG. A tunnel water leakage detection method based on multi-scale edge information fusion[J]. Acta Geodaetica et Cartographica Sinica, 2026, 55(6): 1116-1127.

导出引用管理器 EndNote|Reference Manager|ProCite|BibTeX|RefWorks

链接本文: http://xb.chinasmp.com/CN/10.11947/j.AGCS.2026.20260007

http://xb.chinasmp.com/CN/Y2026/V55/I6/1116

图/表 14

图1

图2

图3

图4

图5

图6

图7

图8

图9

表1

表2

图10

图11

表3

参考文献 34

[1]	赵勇, 田四明, 巩江峰, 等. 截至2025年底中国铁路隧道情况统计分析及中国铁路隧道“十四五”建设成就[J]. 隧道建设(中英文), 2026, 46(4): 877-892.
	ZHAO Yong, TIAN Siming, GONG Jiangfeng, et al. Statistical analysis of railway tunnels in China by the end of 2025 and achievements in the 14th Five-Year Plan period[J]. Tunnel Construction, 2026, 46(4): 877-892.
[2]	KO B, SON Y K, SHIN D, et al. Development of an inspection system for cracks on the lining of concrete tunnels[C]//Proceedings of the 20th ISARC. Eindhoven: [s.n.], 2003: 457-463.
[3]	王如路. 上海轨道交通隧道结构安全性分析[J]. 地下工程与隧道, 2011(4): 37-43.
	WANG Rulu. Structural safety analysis of Shanghai rail transit tunnel[J]. Underground Engineering and Tunnels, 2011(4): 37-43.
[4]	何国华, 刘新根, 陈莹莹, 等. 基于数字图像的隧道表观病害识别方法研究[J]. 重庆交通大学学报(自然科学版), 2019, 38(3): 21-26.
	HE Guohua, LIU Xingen, CHEN Yingying, et al. Apparent tunnel diseases identification based on digital images[J]. Journal of Chongqing Jiaotong University (Natural Science), 2019, 38(3): 21-26.
[5]	HUANG Zhen, FU Helin, ZHANG Jiabing, et al. Structural damage evaluation method for metro shield tunnel[J]. Journal of Performance of Constructed Facilities, 2019, 33(1): 04018097.
[6]	HU Chuanpeng, ZHU Hehua, LI Xiaojun. Detection of tunnel water leakage based on image processing[C]//Proceedings of the 1st International Conference on Information Technology in Geo-Engineering. Shanghai: [s.n.], 2010: 254-262.
[7]	CHA Y J, CHOI W, BÜYÜKÖZTÜRK O. Deep learning-based crack damage detection using convolutional neural networks[J]. Computer-Aided Civil and Infrastructure Engineering, 2017, 32(5): 361-378. DOI: . doi: 10.1111/mice.12263
[8]	薛亚东, 高健, 李宜城, 等. 基于深度学习的地铁隧道衬砌病害检测模型优化[J]. 湖南大学学报(自然科学版), 2020, 47(7): 137-146.
	XUE Yadong, GAO Jian, LI Yicheng, et al. Optimization of shield tunnel lining defect detection model based on deep learning[J]. Journal of Hunan University (Natural Sciences), 2020, 47(7): 137-146.
[9]	REN Shaoqing, HE Kaiming, GIRSHICK R, et al. Faster R-CNN: towards real-time object detection with region proposal networks[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2017, 39(6): 1137-1149.
[10]	MAKANTASIS K, PROTOPAPADAKIS E, DOULAMIS A, et al. Deep convolutional neural networks for efficient vision based tunnel inspection[C]//Proceedings of 2015 IEEE International Conference on Intelligent Computer Communication and Processing. Cluj-Napoca: IEEE, 2015: 335-342.
[11]	MAN Ke, LIU Ruilin, LIU Xiaoli, et al. Water leakage and crack identification in tunnels based on transfer-learning and convolutional neural networks[J]. Water, 2022, 14(9): 1462.
[12]	ZHAO Shuai, ZHANG Dongming, HUANG Hongwei. Deep learning-based image instance segmentation for moisture marks of shield tunnel lining[J]. Tunnelling and Underground Space Technology, 2020, 95: 103156.
[13]	HE Kaiming, GKIOXARI G, DOLLAR P, et al. Mask R-CNN[C]//Proceedings of 2017 IEEE International Conference on Computer Vision. Venice: IEEE, 2017: 2980-2988.
[14]	薛亚东, 贾非, 郭春生, 等. 基于深度学习的盾构隧道渗漏水病害混合样本集构建与精细分割[J]. 应用基础与工程科学学报, 2023, 31(4): 1032-1042.
	XUE Yadong, JIA Fei, GUO Chunsheng, et al. Deep learning-based shield tunnel leakage mixed dataset construction and fine segmentation[J]. Journal of Basic Science and Engineering, 2023, 31(4): 1032-1042.
[15]	GENG Peng, JIA Mengye, REN Xiaoxia. Tunnel lining water leakage image segmentation based on improved BlendMask[J]. Structural Health Monitoring, 2023, 22(2): 865-878.
[16]	GENG Peng, TAN Ziye, LUO Jun, et al. ACPA-Net: atrous channel pyramid attention network for segmentation of leakage in rail tunnel linings[J]. Electronics, 2023, 12(2): 255.
[17]	LIU Shichang, XU Xu, JEON G, et al. Deep learning based water leakage detection for shield tunnel lining[J]. Frontiers of Structural and Civil Engineering, 2024, 18(6): 887-898.
[18]	HAN Lei, CHEN Jiangfan, LI Haobo, et al. Multispectral water leakage detection based on a one-stage anchor-free modality fusion network for metro tunnels[J]. Automation in Construction, 2022, 140: 104345.
[19]	GUO Junyi, FANG Jinhui, SHI Bin, et al. High-sensitivity water leakage detection and localization in tunnels using novel ultra-weak fiber Bragg grating sensing technology[J]. Tunnelling and Underground Space Technology, 2024, 144: 105574.
[20]	YANG Z, LEE W C, CHAN H N, et al. A real-time tunnel surface inspection system using Edge-AI on drone[C]//Proceedings of 2022 IEEE International Conference on Mechatronics and Automation. Guilin: IEEE, 2022: 749-754.
[21]	JIANG Jundi, SHEN Yueqian, WANG Jinhu, et al. GEPT-Net: an efficient geometry enhanced point transformer for shield tunnel leakage segmentation[J]. ISPRS Journal of Photogrammetry and Remote Sensing, 2025, 221: 20-43.
[22]	ZHOU Zhong, ZHANG Junjie, GONG Chenjjie. Automatic detection method of tunnel lining multi-defects via an enhanced you only look once network[J]. Computer-Aided Civil and Infrastructure Engineering, 2022, 37(6): 762-780.
[23]	CHEN Junxin, XU Xu, JEON G, et al. WLR-Net: an improved YOLO-v7 with edge constraints and attention mechanism for water leakage recognition in the tunnel[J]. IEEE Transactions on Emerging Topics in Computational Intelligence, 2024, 8(4): 3105-3116.
[24]	高贤君, 刘振宇, 许磊, 等. 基于CutMix数据增强与多约束损失函数的YOLOv7盾构隧道渗漏水检测[J]. 测绘通报, 2024(7): 105-110.
	GAO Xianjun, LIU Zhenyu, XU Lei, et al. YOLOv7 shield tunnel water leakage detection based on CutMix data augmentation and multi-constraint loss function[J]. Bulletin of Surveying and Mapping, 2024(7): 105-110.
[25]	纪长琦, 郭肇捷, 孙海丽, 等. 基于移动激光扫描的地铁隧道渗漏水定位及快速检测方法[J]. 测绘学报, 2024, 53(6): 1236-1250. DOI: . doi: 10.11947/j.AGCS.2024.20230438
	JI Changqi, GUO Zhaojie, SUN Haili, ZHONG Ruofei. Location and rapid detection method of water leakage in subway tunnels based on mobile laser scanning[J]. Acta Geodaetica et Cartographica Sinica, 2024, 53(6): 1236-1250. DOI: . doi: 10.11947/j.AGCS.2024.20230438
[26]	张振海, 孙岩, 李哲远. 基于MDS-YOLO的轻量级隧道表观病害检测算法[J]. 交通运输工程学报, 25(6): 271-283.
	ZHANG Zhenhai, SUN Yan, LI Zheyuan. Lightweight tunnel surface defect detection algorithm based on MDS-YOLO[J]. Journal of Traffic and Transportation Engineering, 2025, 25(6): 271-283.
[27]	PERONA P, MALIK J. Scale-space and edge detection using anisotropic diffusion[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 1990, 12(7): 629-639.
[28]	PIZER S M, AMBURN E P, AUSTIN J D, et al. Adaptive histogram equalization and its variations[J]. Computer Vision, Graphics, and Image Processing, 1987, 39(3): 355-368.
[29]	YASEEN M. What is YOLOv8: an in-depth exploration of the internal features of the next-generation object detector[EB/OL]. [2025-12-11]. https://arxiv.org/abs/2408.15857.
[30]	XIAO Yao, XU Tingfa, XIN Yu, et al. FBRT-YOLO: faster and better for real-time aerial image detection[EB/OL]. [2025-12-11]. https://arxiv.org/pdf/2504.20670.
[31]	ZHAO Yian, LÜ Wenyu, XU Shangliang, et al. DETRs beat YOLOs on real-time object detection[C]//Proceedings of 2024 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Seattle: IEEE, 2024: 16965-16974.
[32]	KHANAM R, HUSSAIN M. YOLOv11: an overview of the key architectural enhancements[EB/OL]. [2025-12-11]. https://arxiv.org/abs/2410.17725.
[33]	TIAN Yunjie, YE Qixiang, DOERMANN D. YOLOv12: attention-centric real-time object detectors[EB/OL]. [2025-12-11]. https://arxiv.org/abs/2502.12524.
[34]	FENG Y, HUANG J, DU S, et al. Hyper-YOLO: when visual object detection meets hypergraph computation[J]. IEEE transactions on pattern analysis and machine intelligence, 2024, 47(4): 2388-2401.

输入数据	mAP	精确率	召回率
YOLOv8（原始图像）	71.9	70.2	68.2
YOLOv8（增强图片）	73.6	71.9	71.2

方法	mAP	精确率	召回率
YOLOv11	77.9	72.3	75.8
YOLOv12	77.9	74.8	75.4
RT-DETR	77.5	72.0	72.3
FBRT-YOLO	74.5	72.5	70.6
Hyper-YOLO	78.3	73.6	74.3
GLEF-YOLO	82.0	77.0	79.5

模型结构	mAP	精确率	召回率
YOLOv8	73.6	71.9	71.2
YOLOv8+MEIT	80.0	76.8	75.3
YOLOv8+MEIT+GLAFusion	82.0	77.0	79.5