| 14 | 0 | 59 |
| 下载次数 | 被引频次 | 阅读次数 |
为了提升板式橡胶支座的病害检测精度，采用YOLO(You Only Look Once)系列中的YOLOv3、YOLOv4、YOLOR这3种算法针对板式橡胶支座的多种表观病害进行检测并对其检测性能进行对比研究。建立由6787张图片组成的支座病害(BD)数据集，涵盖开裂、锈蚀、脱空、龟裂、垃圾、外鼓和剪切变形7种病害。采用YOLOv3、YOLOv4、YOLOR算法基于BD数据集进行训练，并给出了与之对应的3种检测方法。基于此，对3种检测方法进行检测性能对比实验。研究结果表明：针对板式橡胶支座多表观病害，YOLOv3与YOLOv4检测性能接近，而YOLOR相较YOLOv3和YOLOv4整体召回率(R)和平均精度均值(PmA)提升显著，分别提升了59.7%和92.7%,针对各类病害检测性能也存在不同程度的提升。因此，基于YOLOR的检测方法更适用于板式橡胶支座多种表观病害的检测。
Abstract：To enhance the detection accuracy of apparent defects in laminated bearings, three algorithms, YOLOv3, YOLOv4 and YOLOR from the YOLO(You Only Look Once) series, are used to detect the various apparent defects in laminated bearings, and comparative research is conducted to evaluate their respective detection performance. Initially, a bearing defect dataset(BD dataset) comprising 6787 images is established, including seven types of defects, namely crack, bulging, rust, void, crazing, shear deformation and garbage. Subsequently, the YOLOv3, YOLOv4 and YOLOR are employed for training based on the BD dataset, and three corresponding detection methods are provided. Following this, a comparative experimental analysis of the detection performance of the three detection methods is conducted. The results indicate that for the multiple apparent defects in laminated bearings, the detection performance of YOLOv3 is close to that of YOLOv4, while compared with YOLOv3 and YOLOv4, YOLOR exhibits a significant increase of 59.7% in overall R and92.7% in PmA. Additionally, varying degrees of improvement in the detection performance for different types of defects are observed. Therefore, the detection method based on YOLOR is more suitable for the detection of multiple apparent defects in laminated bearings.
[1] 吴宜峰，徐泓，李爱群，等.桥梁板式橡胶支座运营病害及影响研究综述[J].工程抗震与加固改造，2021,43(5):104-113.WU Yifeng,XU Hong,LI Aiqun,et al.Review on operating disease and effects of bridge plate rubber bearing[J].Earthquake Resistant Engineering and Retrofitting,2021,43(5):104-113.
[2] 鲍跃全，李惠.人工智能时代的土木工程[J].土木工程学报，2019,52(5):1-11.BAO Yuequan,LI Hui.Artificial intelligence for civil engineering[J].China Civil Engineering Journal,2019,52(5):1-11.
[3] LECUN Y,BENGIO Y,HINTON G.Deep learning[J].Nature,2015,521:436-444.
[4] ZHANG Y,HUANG J,CAI F.On bridge surface crack detection based on an improved YOLOv3 algorithm[J].IFAC-PapersOnLine,2020,53(2):8205-8210.
[5] 邹俊志，杨建喜，李昊，等.复杂背景下基于改进YOLOv3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.ZOU Junzhi,YANG Jianxi,LI Hao,et al.Bridge apparent damage detection based on the improved YOLOv3 in complex background[J].Journal of Railway Science and Engineering,2021,18(12):3257-3266.
[6] HAN Q,ZHAO N,XU J.Recognition and location of steel structure surface corrosion based on unmanned aerial vehicle images[J].Journal of Civil Structural Health Monitoring,2021,11(5):1375-1392.
[7] 谌婷婷，魏怡.基于改进YOLOv4的混凝土裂缝检测方法[J].激光杂志，2024,45(1):80-85.SHEN Tingting,WEI Yi.Improved YOLOv4-based concrete crack detection method[J].Laser Journal,2024,45(1):80-85.
[8] YU Z,SHEN Y,SHEN C.A real-time detection approach for bridge cracks based on YOLOv4-FPM[J].Automation in Construction,2021,122:103514.
[9] GARCIA-SANCHEZ D,FERNANDEZ-NAVAMUEL A,ZAMORA S D,et al.Bearing assessment tool for longitudinal bridge performance[J].Journal of Civil Structural Health Monitoring,2020,10(5):1023-1036.
[10] 陈红霞，狄方殿，朱亚洲.连续梁桥支座损伤识别方法[J].土木工程与管理学报，2017,34(4):53-58.CHEN Hongxia,DI Fangdian,ZHU Yazhou.Bearing damage detection of continious bridge[J].Journal of Civil Engineering and Management,2017,34(4):53-58.
[11] 崔弥达.基于图像处理的桥梁支座病害自动识别关键技术研究[D].南京：东南大学，2018.CUI Mida.Research of on key techniques of automatic defect identification for bridge bearings based on image processing[D].Nanjing:Southeast University,2018.
[12] CUI M,WU G,DANG J,et al.Deep learning-based condition assessment for bridge elastomeric bearings[J].Journal of Civil Structural Health Monitoring,2022,12(2):245-261.
[13] MA C F,ZHONG K J,HUANG P M.Application analysis of bridge support safety detection recognition and deep learning image processing technology[C]// International Conference on Transportation and Development,ASCE 2020,2020:1873-1884.
[14] CHEN Z,FANG H,KE X,et al.A new method to identify bridge bearing damage based on radial basis function neural network[J].Earthquakes and Structures,2016,11(5):841-859.
[15] SHI J,DANG J,CUI M,et al.Improvement of damage segmentation based on pixel-level data balance using VGG-Unet[J].Applied Sciences,2021,11(2):518.
[16] REDMON J,FARHADI A.YOLOv3:An incremental improvement[M].San Diego(CA):ArXiv E-prints,2018.
[17] BOCHKOVSKIY A,WANG C Y,LIAO H Y M.YOLOv4:Optimal speed and accuracy of object detection[M].San Diego(CA):ArXiv E-prints,2020.
[18] WANG C Y,YEH I H,LIAO H Y M.You only learn one representation:Unified network for multiple tasks[M].San Diego(CA):ArXiv E-prints,2021.
[19] JTG/T H21—2011,公路桥梁技术状况评定标准[S].JTG/T H21—2011,Standards for technical condition evaluation of highway bridges[S].
[20] JTG 5120—2021,公路桥涵养护规范[S].JTG 5120—2021,Specifications for maintenance of highway bridges and culverts[S].
基本信息:
DOI：
中图分类号:U445.71
引用信息:
[1]鲁德文,李晰.基于YOLO的板式橡胶支座多表观病害检测方法对比研究[J].青岛理工大学学报,2025,46(03):1-8.
基金信息:
国家重点研发计划(2022YFC3801104)