地铁隧道病害检测深度学习模型优化及应用

doi:10.13474/j.cnki.11-2246.2024.0817

测绘通报 ›› 2024, Vol. 0 ›› Issue (8): 96-101.doi: 10.13474/j.cnki.11-2246.2024.0817

• 学术研究 • 上一篇

地铁隧道病害检测深度学习模型优化及应用

尤相骏¹, 赵霞², 龙四春³, 王嘉伟¹, 郑颖², 邝利军⁴

1. 浙江华展研究设计院股份有限公司, 浙江宁波 315000;
2. 北京工商大学计算机与人工智能学院, 北京 102446;
3. 湖南科技大学地球科学与空间信息工程学院, 湖南湘潭 411201;
4. 中国建筑第五工程局有限公司, 广东深圳 518108

收稿日期:2024-03-11 发布日期:2024-09-03
通讯作者: 赵霞。E-mail:zhaox@btbu.edu.cn
作者简介:尤相骏(1978—),男,硕士,高级工程师,主要研究方向为工程测量与智能化测绘。E-mail:karlyou@qq.com
基金资助:
国家自然科学基金(42377453;41877283);湖南省科技创新计划(2021RC4037;2033JJ30235);湖南省自然资源厅科研项目(2021-18);2024年度宁波市“科创甬江2035”关键技术突破计划科研项目

Optimization and application of deep learning model-based subway tunnel defect detection

YOU Xiangjun¹, ZHAO Xia², LONG Sichun³, WANG Jiawei¹, ZHENG Ying², KUANG Lijun⁴

1. Zhejiang Huazhan Institute of Design Co., Ltd., Ningbo 315000, China;
2. School of Artificial Intelligence, Beijing University of Technology and Business, Beijing 102446, China;
3. School of Earth Science and Space Information Engineering, Hunan University of Science and Technology, Xiangtan 411201, China;
4. China Construction Fifth Engineering Bureau Co., Ltd., Shenzhen 518108, China

Received:2024-03-11 Published:2024-09-03

摘要/Abstract

摘要： 本文针对地铁隧道中渗漏水、裂纹裂缝、结构抹灰开裂及剥落掉块4种常见病害,研究了基于激光雷达扫描点云数据和深度学习的地铁隧道病害检测方法。首先,引入ACmix注意力模块,使网络兼顾全局特征和局部特征,提升对裂纹、裂缝等小目标的检测效果;然后,优化回归损失函数,提高收敛平稳度和回归精度,降低检测误差;最后,实现正射投影图像预处理、批量检测、结果融合及检测结果报表的一体化生成,提高大尺度正射投影图的病害检测率。试验结果表明,在选取IoU阈值为0.5的条件下,改进后的YOLOv8算法在隧道病害测试检测中正确率由90.65%提升至91.18%,基本实现了基于激光雷达扫描的地铁隧道4类常见病害的智能检测,并在实际隧道运维工程中得到成功应用。

关键词: 深度学习, 模型优化, 检测方法, 隧道病害

Abstract: Aiming at the four common defects of subway tunnel, such as leakage, crack, structural plaster cracking and spalling, a defect detection method of subway tunnel based on laser radar scanning point cloud data and deep learning is studied.Firstly,the ACmix attention module is introduced into the YOLOv8 model to make the network take into account both global and local features, and improve the detection effect of small targets such as cracks and cracks.Then,the regression loss function is optimized, the convergence stability and regression accuracy are improved, and the detection error is reduced. Finally,the complete process of orthographic projection image preprocessing, batch detection and result fusion, and report generation of detection results is realized, and the defect detection of large-scale orthographic projection is efficiently realized. The experimental results show that under the condition that the IoU threshold is 0.5, the mAP of the improved YOLOv8 algorithm on the tunnel defect test set increases from 90.65% to 91.18%, and the AP of cracks increases from 77.89% to 78.70%. The intelligent detection of four common defects of subway tunnel based on LiDAR scanning is solved, and has been successfully applied in actual tunnel operation and maintenance engineering.

Key words: deep learning, model optimization, inspection method, tunnel defects

中图分类号:

P237

尤相骏, 赵霞, 龙四春, 王嘉伟, 郑颖, 邝利军. 地铁隧道病害检测深度学习模型优化及应用[J]. 测绘通报, 2024, 0(8): 96-101.

YOU Xiangjun, ZHAO Xia, LONG Sichun, WANG Jiawei, ZHENG Ying, KUANG Lijun. Optimization and application of deep learning model-based subway tunnel defect detection[J]. Bulletin of Surveying and Mapping, 2024, 0(8): 96-101.

参考文献

[1] 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.
[2] 路耀邦,刘永胜,樊晓东.地铁隧道结构表观病害快速检测方法与应用[J].隧道建设(中英文),2021,41(S2):655-663.
[3] 刘德军,仲飞,黄宏伟,等.运营隧道衬砌病害诊治的现状与发展[J].中国公路学报,2021,34(11): 178-199.
[4] 史玉峰,张俊,张迎亚.基于地面三维激光扫描技术的隧道安全监测[J].东南大学学报(自然科学版),2013,43(S2):246-249.
[5] MUDULIi P R,PATI U C.A novel technique for wall crack detection using image fusion[C]//Proceedings of 2013 International Conference on Computer Communication and Informatics.Coibatore:IEEE,2013.
[6] IYER S,SINHA S K,TITTMANN B R,et al.Ultrasonic signal processing methods for detection of defects in concrete pipes[J].Automation in Construction,2012,22:135-148.
[7] CHA Y J,CHOI W,BüYüKÖZTÖRK O.Deep learning-based crack damage detection using convolutional neural networks[J].Computer-Aided Civil and Infrastructure Engineering,2017,32(5): 361-378.
[8] XUE Y,LI Y.A fast detection method via region-based fully convolutional neural networks for shield tunnel lining defects[J].Computer-Aided Civil and Infrastructure Engineering,2018,33(8): 638-654.
[9] 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.
[10] 柴雪松,朱兴永,李健超,等.基于深度卷积神经网络的隧道衬砌裂缝识别算法[J].铁道建筑,2018,58(6): 60-65.
[11] ZHOU Z,ZHANG J,GONG C.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.
[12] 杨锋,丁之桐,邢蒙蒙,等.深度学习的目标检测算法改进综述[J].计算机工程与应用,2023,59(11): 1-15.
[13] ZHOU Z,ZHANG J,GONG C,et al.Automatic tunnel lining crack detection via deep learning with generative adversarial network-based data augmentationt[J].Underground Space,2023,9(1):140-154.
[14] CARION N,MASSA F,SYNNAEVE G,et al.End-to-end object detection with transformers[C]//Proceedings of the 16th European Conference Computer Vision.Glasgow:Springer International Publishing,2020: 213-229.
[15] TERVEN J,CORDOVA-ESPARZA D.A comprehensive review of YOLO: from YOLOv1 to YOLOv8 and beyond[J/OL].arXiv.[2024-02-19].https://doi.org/10.48550/arXiv.2304.00501.
[16] PAN X,GE C,LU R,et al.On the integration of self-attention and convolution[C]//Proceedings of 2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition.Orleans:IEEE,2022: 815-825.
[17] ZHENG Z,WANG P,REN D,et al.Enhancing geometric factors in model learning and inference for object detection and instance segmentation[J].IEEE Transactions on Cybernetics,2021,52(8): 8574-8586.
[18] ZHENG Z,WWANG P,LIU W,et al.Distance-IoU loss: faster and better learning for bounding box regression[C]//Proceedings of 2020 AAAI Conference on Artificial Intelligence.2020: 12993-13000.
[19] GEVORGYAN Z.SIoU loss: more powerful learning for bounding box regression[J/OL].arXiv.[2024-02-19].https://doi.org/10.48550/arXiv.2205.12740.
[20] CRAIG J C.A confusion matrix for tactually presented letters[J].Perception & Psychophysics,1979.26(5): 409-411.

地铁隧道病害检测深度学习模型优化及应用

Optimization and application of deep learning model-based subway tunnel defect detection

PDF

可视化

摘要/Abstract

引用本文

使用本文

参考文献

相关文章 15

编辑推荐

Metrics

本文评价

[1]	石一剑, 谭海, 钟旭辉. 改进YOLOv5的国产光学影像辐射异常检测方法[J]. 测绘通报, 2024, 0(8): 60-65,72.
[2]	杨晓峰, 魏浩翰, 张强, 向云飞. 融合孤立森林和深度学习的GNSS-IR土壤湿度反演[J]. 测绘通报, 2024, 0(7): 88-94.
[3]	田亚林, 连增增, 王鹏辉, 王孟奇, 陆力. 基于KF-LSTM的UWB室内定位算法[J]. 测绘通报, 2024, 0(7): 95-99,151.
[4]	刘兵, 刘宇, 金凤学, 邹一波, 葛艳, 赵林林. 针对浒苔目标检测的全局背景强化的位置蒸馏方法[J]. 测绘通报, 2024, 0(6): 19-23.
[5]	高墨通, 杨维芳, 刘祖昱, 曹小双, 张瑞琪, 侯宇豪. 结合卷积神经网络和注意力机制的LSTM采空区地表沉降预测方法[J]. 测绘通报, 2024, 0(6): 53-58,170.
[6]	钟青岑, 吴晨昊, 向隆刚, 姚鹏. 顾及路网约束的深度地图匹配方法[J]. 测绘通报, 2024, 0(6): 96-102,133.
[7]	尹泽中, 李功权. 三维场叠加视频流的电子围栏越界检测方法[J]. 测绘通报, 2024, 0(6): 103-108.
[8]	王丽霞, 喜文飞, 史正涛, 赵子龙, 钱堂慧, 赵磊, 马奕捷. 深度学习滑坡识别算法中样本不平衡问题的研究[J]. 测绘通报, 2024, 0(5): 12-18.
[9]	姚楚羡, 蔡皓楠, 张远波, 唐可懿, 詹璐, 周宝定. 基于轻量化车载设备的道路病害检测方法[J]. 测绘通报, 2024, 0(5): 147-150.
[10]	张楠, 张蕴灵. 基于深度学习模型的复杂困难路网提取方法[J]. 测绘通报, 2024, 0(5): 155-159.
[11]	林泊锟, 丁勇, 李登华. 一种基于深度学习与图像局部特征提取的边坡异常监测技术[J]. 测绘通报, 2024, 0(4): 23-28.
[12]	庞敏. 基于深度学习的国产卫片图斑提取[J]. 测绘通报, 2024, 0(4): 124-128,134.
[13]	陈昶, 张岩, 李坤衡, 杨润泽, 张俊彬, 梁彦荣. 基于高分七号影像自动提取东北黑土区侵蚀沟的方法[J]. 测绘通报, 2024, 0(3): 1-7.
[14]	张兰兰, 王红雷. 基于深度学习的耕地非农化遥感监测与时空分析——以开阳县为例[J]. 测绘通报, 2024, 0(3): 13-18.
[15]	王鸿雁, 车向红, 徐辛超, 徐胜华, 李洪胜. 利用DeepLabv3+模型提取分析街景图像绿视率——以北京三环内为例[J]. 测绘通报, 2024, 0(3): 88-94.