Please wait a minute...
浙江大学学报(工学版)
浙江大学学报(工学版)  2022, Vol. 56 Issue (1): 92-99    DOI: 10.3785/j.issn.1008-973X.2022.01.010
土木工程、水利工程     
基于深度学习的公路隧道衬砌病害识别方法
任松(),朱倩雯,涂歆玥,邓超,王小书
重庆大学 煤矿灾害动力学与控制国家重点实验室, 重庆 400044
Lining disease identification of highway tunnel based on deep learning
Song REN(),Qian-wen ZHU,Xin-yue TU,Chao DENG,Xiao-shu WANG
State Key Laboratory of Coal Mine Disaster Dynamics and Control, Chongqing University, Chongqing 400044, China
 全文: PDF(1492 KB)   HTML
摘要：

针对与日俱增的隧道养护需求，为了节约时间与人力成本，提出基于卷积神经网络的公路隧道衬砌病害检测方法. 利用自主研制的隧道智能快速检测车采集24条隧道衬砌的图像，构建超过20 000张病害图像的高质量数据集. 结合隧道衬砌病害的成因及特点，分别构建单阶段SSD模型和两阶段R-FCN模型在自制的数据集上训练，对检测结果进行对比分析，提出离线式隧道衬砌病害检测方案. 试验结果表明，SSD模型的识别准确率为98%，总的平均精度均值（mAP）为72%，检测速度较快，适用于隧道的快速诊断. R-FCN模型的识别准确率为85%，总的mAP达到91%，检测精度较高，适用于隧道病害的后期处理. 利用这2种检测模型均可以提升检测效率和精度.
关键词： 深度学习;  卷积神经网络;  隧道衬砌;  隧道病害检测;  裂缝;  渗漏水    
Abstract:

A highway tunnel lining disease detection method based on convolutional neural network was proposed aiming at the ever-increasing demand for tunnel maintenance in order to save time and labor costs. The self-developed intelligent rapid detection vehicle for tunnels was used to collect 24 tunnel lining images. A high-quality data set of more than 20000 disease images was constructed. Then single-stage SSD (single shot multiBox detector) models and two-stage R-FCN (region-based fully convolutional networks) models were constructed on a self-made data set combining the causes and characteristics of tunnel lining diseases. The detection results were compared and analyzed, and an offline tunnel lining disease detection scheme was proposed. The experimental results showed that the identification accuracy rate of SSD model was 98%, the total mean average precision (mAP) was 72%, and the detection speed was fast. The SSD model is suitable for rapid diagnosis of tunnels. The identification accuracy rate of R-FCN model was 85%, the total mAP value reached 91%, and the detection accuracy was high. The R-FCN model is suitable for the post-treatment of tunnel diseases. Using these two detection models can improve detection efficiency and accuracy.
Key words: deep learning    convolutional neural network    tunnel lining    tunnel disease detection    crack    leakage
收稿日期: 2021-02-13 出版日期: 2022-01-05
CLC:  U 456  
基金资助: 国家自然科学基金资助项目(51774057，52074048)；重庆市研究生科研创新资助项目(CYS18020)
作者简介: 任松(1975—)，男，教授，博导，从事岩土工程监测及可靠性分析、岩层稳定性及控制、盐层油气储库造腔理论及工艺、煤与瓦斯突出机理及控制研究. orcid.org/0000-0002-1822-5991. E-mail： rs_rwx@cqu.edu.cn
服务　　
把本文推荐给朋友
加入引用管理器
E-mail Alert
作者相关文章　　
任松
朱倩雯
涂歆玥
邓超
王小书

引用本文:

任松,朱倩雯,涂歆玥,邓超,王小书. 基于深度学习的公路隧道衬砌病害识别方法[J]. 浙江大学学报(工学版), 2022, 56(1): 92-99.

Song REN,Qian-wen ZHU,Xin-yue TU,Chao DENG,Xiao-shu WANG. Lining disease identification of highway tunnel based on deep learning. Journal of ZheJiang University (Engineering Science), 2022, 56(1): 92-99.

链接本文:

https://www.zjujournals.com/eng/CN/10.3785/j.issn.1008-973X.2022.01.010    或    https://www.zjujournals.com/eng/CN/Y2022/V56/I1/92

图 1  自制的隧道及路面智能快速检测车
图 2  LabelImg软件隧道病害图像标注界面
隧道病害种类 Ntr Nv Nt N
A 15419 2771 1949 20139
B 1991 345 257 2593
S 17410 3116 2206 22732
表 1  隧道病害数据集
图 3  SSD网络模型结构[20]
图 4  基于SSD网络模型的隧道病害检测方案
图 5  基于R-FCN网络模型的隧道病害检测方案
图 6  提取候选框网络的结构
图 7  bottleneck网络结构
图 8  SSD-Inception V2模型总损失变化曲线
图 9  R-FCN模型总损失变化曲线
模型 病害类型 P R AP mAP t/h V
SSD-Mobilenet 裂缝 0.996 0.402 0.411 0.364 $34\dfrac{1}{{6}}$ 2.06
SSD-Mobilenet 渗水 1 0.311 0.316
SSD-Inception V2 裂缝 0.996 0.757 0.769 0.728 $48\dfrac{1}{{3}} $ 3.17
SSD-Inception V2 渗水 0.988 0.673 0.687
R-FCN 裂缝 0.935 0.981 0.983 0.910 $87\dfrac{3}{{4}} $ 10.30
R-FCN 渗水 0.835 0.805 0.833
Faster R-CNN 裂缝 0.796 0.994 0.994 0.890 $84\dfrac{1}{{2}} $ 9.42
Faster R-CNN 渗水 0.886 0.790 0.791
表 2  不同卷积神经网络的隧道病害检测结果的对比
图 10  渗水检测可视化识别结果对比
图 11  裂缝检测可视化识别结果对比
图 12  隧道衬砌病害的检测流程
图 13  隧道采集图像处理前、后的对比
1 黄宏伟, 李庆桐 基于深度学习的盾构隧道渗漏水病害图像识别[J]. 岩石力学与工程学报, 2017, 36 (12): 2861- 2871
HUANG Hong-wei, LI Qing-tong Image recognition for water leakage in shield tunnel based on deep learning[J]. Chinese Journal of Rock Mechanics and Engineering, 2017, 36 (12): 2861- 2871
2 牛晓博, 何雪婷 铁路运营隧道病害现状及检测评估[J]. 科技经济导刊, 2019, 27 (16): 86
NIU Xiao-bo, HE Xue-ting Current damage situation of railway operation tunnels and their inspection and evaluation[J]. Technology and Economic Guide, 2019, 27 (16): 86
3 林亮伦, 杜清超, 徐志武 重庆某隧道病害调查分析与治理对策[J]. 重庆建筑, 2018, 17 (8): 50- 53
LIN Liang-lun, DU Qing-chao, XU Zhi-wu Investigation analysis and treatments of tunnel diseases of a tunnel in Chongqing[J]. Chongqing Architecture, 2018, 17 (8): 50- 53
doi: 10.3969/j.issn.1671-9107.2018.08.50
4 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, 25 (4): 37- 43
WANG Ru-lu Structural safety analysis of Shanghai rail transit tunnel[J]. Underground Engineering and Tunnel, 2011, 25 (4): 37- 43
5 刘海京, 夏才初, 朱合华, 等 隧道病害研究现状与进展[J]. 地下空间与工程学报, 2007, 3 (5): 947- 953
LIU Hai-jing, XIA Cai-chu, ZHU He-hua, et al Studies on tunnel damage[J]. Chinese Journal of Underground Space and Engineering, 2007, 3 (5): 947- 953
6 赵永贵, 刘浩, 孙宇, 等. 隧道超前预报与病害诊断技术[C]// 全国公路工程关键技术与管理信息化学术会议. 长春: 中国公路学会, 2003: 282-288.
ZHAO Yong-gui, LIU Hao, SUN Yu, et al. Tunnel advanced forecast and disease diagnosis technology [C]// Tunnel Advance Forecast and Disease Diagnosis Technology National Conference on Key Technology and Management Informationization of Highway Engineering. Changchun: China Highway Society, 2003: 282-288.
7 杨正华, 张文波, 王卫东, 等 浅析地震波法用于隧道病害的诊断与预测[J]. 灾害学, 2004, 19 (1): 27- 30
YANG Zheng-hua, ZHANG Wen-bo, WANG Wei-dong, et al Analysis on tunnel hazard diagnosis and forecast by seismic wave[J]. Journal of Catastrophology, 2004, 19 (1): 27- 30
doi: 10.3969/j.issn.1000-811X.2004.01.006
8 徐宏武 渝湘高速下塘口一号隧道突水病害诊断及处治[J]. 地下空间与工程学报, 2010, 6 (4): 845- 849
XU Hong-wu Catastrophology gushing water diagnosis and treatment of Yuxiang Speedway on Xiaotangkou No. 1 Tunnel[J]. Chinese Journal of Underground Space and Engineering, 2010, 6 (4): 845- 849
9 MAKANTASIS K, PROTOPAPADAKIS E, DOULAMIS A, et al. Deep convolutional neural networks for efficient vision based tunnel inspection [C]// Proceedings of the 11th International Conference on Intelligent Computer Communication and Processing. Cluj-Napoca: IEEE, 2015: 335-342.
10 PROTOPAPADAKIS E, DOULAMIS N. Image based approaches for tunnels defects recognition via robotic inspectors [C]// Proceedings of the 11th International Symposium on Visual Computing. Las Vegas: Springer, 2015: 706-716.
11 ZHANG L, YANG F, ZHANG Y D, et al. Road crack detection using deep Convolutional neural network [C]// Proceedings of the 23rd International Conference on Image Processing. Phoenix: IEEE, 2016: 3708-3712.
12 CHA Y J, CHOI W, BUYUKOZTURK O Deep learning-based crack damage detection using convolutional neural networks[J]. Computer-Aided Civil and Infrastructure Engineering, 2017, 32 (5): 361- 378
doi: 10.1111/mice.12263
13 汤一平, 胡克钢, 袁公萍 基于全景图像CNN的隧道病害自动识别方法[J]. 计算机科学, 2017, 4 (11A): 207- 211
TANG Yi-ping, HU Ke-gang, YUAN Gong-ping Automatic recognition method of tunnel disease based on convolutional neural network for panoramic images[J]. Computer Science, 2017, 4 (11A): 207- 211
doi: 10.11896/j.issn.1002-137X.2017.11A.043
14 胡利娜 一种基于深度学习的隧道衬砌病害检测技术[J]. 山西电子技术, 2019, 47 (5): 38- 40
HU Li-na Tunnel lining disease detection technology based on deep learning[J]. Shanxi Electronic Technology, 2019, 47 (5): 38- 40
doi: 10.3969/j.issn.1674-4578.2019.05.012
15 薛亚东, 李宜城 基于深度学习的盾构隧道衬砌病害识别方法[J]. 湖南大学学报:自然科学版, 2018, 45 (3): 100- 109
XUE Ya-dong, LI Yi-cheng A method of disease recognition for shield tunnel lining based on deep learning[J]. Journal of Hunan University: Natural Sciences, 2018, 45 (3): 100- 109
16 LIU W, ANGUELOV D, ERHAN D, et al. Ssd: single shot multibox detector [C]// European conference on computer vision. Cham: Springer, 2016: 21-37.
17 FU C Y, LIU W, RANGA A, et al. DSSD: deconvolutional single shot detector [EB/OL]. [2017-01-23]. https://arxiv.org/abs/1701.06659.
18 JEONG J, PARK H, KAWK N. Enhancement of SSD by concatenating feature maps for object detection [EB/OL]. [2018-05-17]. https://arxiv.org/abs/1705.09587.
19 LI Z, ZHOU F. FSSD: feature fusion single shot multibox detector [EB/OL]. [2017-05-26]. https://arxiv.org/abs/1712.00960v1.
20 卢健, 何金鑫, 李哲, 等 基于深度学习的目标检测综述[J]. 电光与控制, 2020, 27 (5): 56- 63
LU Jian, HE Jin-xin, LI Zhe, et al A survey of target detection based on deep learning[J]. Electronics Optics and Control, 2020, 27 (5): 56- 63
doi: 10.3969/j.issn.1671-637X.2020.05.012
21 IOFFE S, SZEGEDY C. Batch normalization: accelerating deep network training by reducing internal covariate shift [EB/OL]. [2015-03-02]. https://arxiv.org/abs/1502.03167.
22 SZEGEDY C, LIU W, JIA YQ, et al. Going deeper with convolutions [C]// Proceedings of 2015 IEEE Conference on Computer Vision and Pattern Recognition. Boston: IEEE, 2015: 1-9.
23 HOWARD A G, ZHU M, CHEN B, et al. MobileNets: efficient convolutional neural networks for mobile vision applications [EB/OL]. [2017-04-17]. https://arxiv.org/abs/1704.04861.
24 DAI Ji-feng, LI Yi, HE Kai-ming, et al. R-FCN: object detection via region-based fully convolutional networks [C]// Proceedings of the 30th International Conference on Neural Information Processing Systems. Barcelona: ACM, 2016: 379-387.
[1] 何立,庞善民. 结合年龄监督和人脸先验的语音-人脸图像重建[J]. 浙江大学学报(工学版), 2022, 56(5): 1006-1016.
[2] 张雪芹,李天任. 基于Cycle-GAN和改进DPN网络的乳腺癌病理图像分类[J]. 浙江大学学报(工学版), 2022, 56(4): 727-735.
[3] 王云灏,孙铭会,辛毅,张博宣. 基于压电薄膜传感器的机器人触觉识别系统[J]. 浙江大学学报(工学版), 2022, 56(4): 702-710.
[4] 褚晶辉,史李栋,井佩光,吕卫. 适用于目标检测的上下文感知知识蒸馏网络[J]. 浙江大学学报(工学版), 2022, 56(3): 503-509.
[5] 程若然,赵晓丽,周浩军,叶翰辰. 基于深度学习的中文字体风格转换研究综述[J]. 浙江大学学报(工学版), 2022, 56(3): 510-519, 530.
[6] 温佩芝,陈君谋,肖雁南,温雅媛,黄文明. 基于生成式对抗网络和多级小波包卷积网络的水下图像增强算法[J]. 浙江大学学报(工学版), 2022, 56(2): 213-224.
[7] 陈彤,郭剑锋,韩心中,谢学立,席建祥. 基于生成对抗模型的可见光-红外图像匹配方法[J]. 浙江大学学报(工学版), 2022, 56(1): 63-74.
[8] 刘兴,余建波. 注意力卷积GRU自编码器及其在工业过程监控的应用[J]. 浙江大学学报(工学版), 2021, 55(9): 1643-1651.
[9] 陈雪云,黄小巧,谢丽. 基于多尺度条件生成对抗网络血细胞图像分类检测方法[J]. 浙江大学学报(工学版), 2021, 55(9): 1772-1781.
[10] 刘嘉诚,冀俊忠. 基于宽度学习系统的fMRI数据分类方法[J]. 浙江大学学报(工学版), 2021, 55(7): 1270-1278.
[11] 金立生,华强,郭柏苍,谢宪毅,闫福刚,武波涛. 基于优化DeepSort的前方车辆多目标跟踪[J]. 浙江大学学报(工学版), 2021, 55(6): 1056-1064.
[12] 许佳辉,王敬昌,陈岭,吴勇. 基于图神经网络的地表水水质预测模型[J]. 浙江大学学报(工学版), 2021, 55(4): 601-607.
[13] 王虹力,郭斌,刘思聪,刘佳琪,仵允港,於志文. 边端融合的终端情境自适应深度感知模型[J]. 浙江大学学报(工学版), 2021, 55(4): 626-638.
[14] 张腾,蒋鑫龙,陈益强,陈前,米涛免,陈彪. 基于腕部姿态的帕金森病用药后开-关期检测[J]. 浙江大学学报(工学版), 2021, 55(4): 639-647.
[15] 郑英杰,吴松荣,韦若禹,涂振威,廖进,刘东. 基于目标图像FCM算法的地铁定位点匹配及误报排除方法[J]. 浙江大学学报(工学版), 2021, 55(3): 586-593.