基于改进YOLO模型的钢筋混凝土构件震后损伤检测

陈梓潇; 宋成浩; 胡晓斌

doi:10.3724/j.gyjzG25071002

基于改进YOLO模型的钢筋混凝土构件震后损伤检测

doi: 10.3724/j.gyjzG25071002

武汉大学土木建筑工程学院, 武汉 430072

基金项目:

国家自然科学基金项目（51578429）。

详细信息

作者简介:
陈梓潇，硕士研究生，主要从事结构健康监测研究，chenzixiao@whu.edu.cn。

通讯作者:
胡晓斌，教授，博士生导师，主要从事结构健康监测研究，newhxb@126.com。

计量
- 文章访问数: 60
- HTML全文浏览量: 11
- PDF下载量: 7
- 被引次数: 0
出版历程
- 收稿日期: 2025-07-10
- 网络出版日期: 2025-09-12

Post-Earthquake Damage Detection of Reinforced Concrete Members Based on the Improved YOLO Model

School of Civil Engineering, Wuhan University, Wuhan 430072, China

摘要

摘要: 首先开展了钢筋混凝土（RC）柱抗震试验，在此基础上考虑外部数据，建立了RC构件震后损伤数据集。然后采用FasterNet网络代替YOLOv5模型的骨干网络，并在Neck网络中引入C3Ghost模块和GhostConv，提出了一种轻量级模型FG-YOLOv5。对该模型进行了训练和测试，并开展了消融试验，最后将其部署至智能手机上，实现了RC构件震后损伤的快速检测。结果表明：相对于普通卷积，部分卷积及GhostConv可以大大地降低计算量；引入FasterNet网络、C3Ghost模块及Ghostconv对YOLOv5模型进行轻量化，可在检测精度提高的情况下，大大降低模型大小及计算量；该研究提出的FG-YOLOv5模型，可方便地部署在手机上，实现RC构件震后损伤快速检测。
- 钢筋混凝土构件 /
- 震后损伤 /
- 快速检测 /
- YOLO模型 /
- 轻量化
Abstract: In this paper， seismic tests of reinforced concrete （RC） column were conducted， based on which the post-earthquake damage dataset of RC member was established considering the additional data obtained externally. A lightweight model， i.e. FG-YOLOv5， was then proposed by replacing the backbone of YOLOv5 model with the FasterNet network and introducing the C3Ghost module and GhostConv in the neck of YOLOv5 model. Based on the dataset， the FG-YOLOv5 model was trained and tested and the ablation test was also carried out. Finally， the model was deployed on a smartphone to achieve rapid post-earthquake damage detection of RC members. The results showed that， compared to the conventional convolution， the partial convolution and GhostConv can greatly reduce the computational cost. By introducing the FasterNet， C3Ghost module and Ghostconv into the YOLO v5 model， greatly smaller model size and computation with a bit higher detection accuracy can be achieved. The FG-YOLOv5 model proposed in this paper can be conveniently deployed on mobile phones for rapid post-earthquake damage detection of RC members.
- reinforced concrete member /
- post-earthquake damage /
- fast detection /
- YOLO model /
- lightweight

HTML全文

参考文献(27)

[1]	周传林. 图像处理技术在混凝土桥梁裂缝检测中的应用研究［J］. 筑路机械与施工机化，2014，31（2）:74-77，80.
[2]	MARINIELLO G，PASTORE T，MENNA C，et al. Structural damage detection and localization using decision tree ensemble and vibration data［J］. Computer-Aided Civil and Infrastructure Engineering，2021，36（9）:1129-1149.
[3]	AMEZQUITA-Sanchez J P，ADELI H. Synchrosqueezed wavelet transform-fractality model for locating，detecting，and quantifying damage in smart highrise building structures［J］. Smart Materials and Structures，2015，24（6），065034.
[4]	PEREZ-Ramirez C A，AMEZQUITA-Sanchez J P，Adeli H，et al. New methodology for modal parameters identification of smart civil structures using ambient vibrations and synchrosqueezed wavelet transform［J］. Engineering Applications of Artificial Intelligence，2016，48:1-12.
[5]	周飞燕，金林鹏，董军. 卷积神经网络研究综述［J］. 计算机学报，2017，40（6）:1229-1251.
[6]	杨国俊，齐亚辉，石秀名. 基于数字图像技术的桥梁裂缝检测综述［J］. 吉林大学学报（工学版），2024，54（2）:313-332.
[7]	REDMON J，DIVVALA S，GIRSHICK R，et al. You only look once:Unified，real-time object detection［C］//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition，F. 2016.
[8]	WEI L，DRAGOMIR A，DUMITRU E，et al. SSD:single shot multibox detector［J］. Springer，Cham，2016. DOI: 10.1007/978-3-319-46448-0_2.
[9]	REDMON J，FARHADI A. YOLOv3:an incremental improvement［J］. arXive-prints，2018. DOI: 10.48550/arXiv.1804.02767.
[10]	BOCHKOVSKIY A，WANG C Y，LIAO H Y M. YOLOv4:optimal speed and accuracy of object detection［J］. 2020. DOI: 10.48550/arXiv.2004.10934.
[11]	GLENN:ultralytics/YOLOv5:v5.0（2020）［EB/OL］. https://github.com/ultralytics/YOLOv5
[12]	ZHANG Y，HUANG J，CAI F. On bridge surface crack detection based on an improved YOLO v3 algorithm［J］. IFAC-Papers OnLine，2020，53（2）:8205-8210.
[13]	邹俊志，杨建喜，李昊，等. 复杂背景下基于改进YOLO v3算法的桥梁表观病害识别［J］. 铁道科学与工程学报，2021，18（12）:3257-3266.
[14]	ZOU D J，ZHANG M，BAI Z L，et al. Multicategory damage detection and safety assessment of post-earthquake reinforced concrete structures using deep learning［J］. Computer-Aided Civil and Infrastructure Engineering，2022，37（9）:1188-1204.
[15]	CUI X N，WANG Q C，LI S，et al. Deep learning for intelligent identification of concrete wind-erosion damage［J］. Automation in Construction，2022，141，104381.
[16]	WU P，LIU A，FU J，et al. Autonomous surface crack identification of concrete structures based on an improved one-stage object detection algorithm［J］. Engineering Structures，2022，272，114962.
[17]	LIU Y，ZHOU T，XU J Y，et al. Rotating target detection method of concrete bridge crack based on YOLO v5［J］. Applied Sciences-Basel，2023，13（20）. DOI: 10.3390/app132011118.
[18]	HOWARD A G，ZHU M，CHEN B，et al. Mobilenets:Efficient convolutional neural networks for mobile vision applications［J］. arXiv preprint arXiv:170404861，2017.
[19]	中华人民共和国住房和城乡建设部. 建筑抗震试验规程:JGJ/T 101—2015［S］. 北京:中国建筑工业出版社，2015.
[20]	CATLIN A C，HEWANADUNGODAGE C，PUJOL S，et al. A Cyberplatform for sharing scientific research data at DataCenterHub［J］. Computing in Science and Engineering，2018，20（3）:49-70.
[21]	YANG F，ZHANG L，YU S，et al. Feature pyramid and hierarchical boosting network for pavement crack detection［J］. IEEE，2020（4）. DOI: 10.1109/TITS.2019.2910595.
[22]	MARTIN M，SAGNIK M，SREENIVAS M，et al. Meta-learning convolutional neural architectures for multi-target concrete defect classification with the concrete defect bridge image dataset［C］// 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition（CVPR）. 2019:11188-11197.
[23]	LAUREN S. Human Signal/ label Img（2022）［EB/OL］. https://github.com/tzutalin/labelImg
[24]	ZHU X K，LYU S C，WANG X，et al. TPH-YOLOv5:Improved YOLOv5 based on transformer prediction head for object detection on drone-captured scenarios［C］// 2021 IEEE/CVF International Conference on Computer Vision Workshops（ICCVW 2021）. 2021:2778-2788.
[25]	CHEN J R，KAO S H，HE H，et al. Run，Don't walk:chasing higher FLOPS for faster neural networks［C］// 2023 IEEE/CVF Conference on Computer Vision and Pattern Recognition（CVPR）. 2023:12021-12031.
[26]	HAN K，WANG Y H，TIAN Q，et al. GhostNet:more features from cheap operations:2020［C］// IEEE/CVF Conference on Computer Vision and Pattern Recognition（CVPR）. 2020:1577-1586.
[27]	Android Studio Giraffe 2022.3，Google［EB/OL］. https://developer.android.google.cn，20/ 02/ 23.