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 (5): 901-908    DOI: 10.3785/j.issn.1008-973X.2022.05.007
土木工程     
基于自注意力机制的桥梁螺栓检测算法
鞠晓臣1(),赵欣欣1,*(),钱胜胜2
1. 中国铁道科学研究院集团有限公司 铁道建筑研究所，北京 100081
2. 中国科学院 自动化研究所，北京 100190
Self-attention mechanism based bridge bolt detection algorithm
Xiao-chen JU1(),Xin-xin ZHAO1,*(),Sheng-sheng QIAN2
1. Railway Engineering Research Institute, China Academy of Railway Sciences Corporation Limited, Beijing 100081, China
2. Institute of Automation, Chinese Academy of Sciences, Beijing 100190, China
 全文: PDF(955 KB)   HTML
摘要：

基于构建的真实桥梁螺栓场景数据集，提出基于自注意力机制与中心点回归（SACPR）的螺栓检测算法. 构建基于真实场景的高质量桥梁螺栓场景数据集，并针对数据不均衡、多样性不够的问题，使用数据增强方法进行数据扩充，从而获得更高的分类精度. 采用基于深度学习框架的SACPR算法检测不同场景下的螺栓，并进行标示. 对螺栓检测准确率进行验证实验，验证SACP算法的有效性. 将试验结果与YOLOv3、Faster-RCNN、RetinaNet这3种算法结果进行对比，发现3种检测方法的识别精度分别为80.56 %、87.71%、93.89%，而所提出的SACPR 算法的识别精度为93.91%，明显优于YOLOv3算法和Faster-RCNN算法；虽然SACPR算法与RetinaNet算法的识别精度较接近，但前者的检测速度是后者的5.6倍.
关键词： 桥梁;  图像识别;  SACPR;  螺栓检测;  自注意力机制    
Abstract:

A bolt detection model algorithm based on self attention mechanism and center point regression (SACPR) was proposed based on the real bridge bolt scene data set. Firstly, a high-quality bridge bolt scene data set based on the real scene was constructed, and for the problems of data imbalance and insufficient diversity, data enhancement method was used to expand the data, so as to obtain higher classification accuracy. Secondly, SACPR model algorithm based on deep learning framework was used to detect bolts in different scenes, and label them. Finally, the validity of the proposed method was verified by the verification experiment of bolt detection accuracy. Comparison was conducted with the results of YOLOv3, Faster-RCNN and RetinaNet, and results showed that the recognition accuracy of the three detection methods was 80.56%, 87.71% and 93.89% respectively, while the recognition accuracy of SACPR model method was 93.91%. The accuracy of SACPR model method was obviously better than that of YOLOv3 model algorithm and Faster-RCNN model algorithm. Although the recognition accuracy was almost the same as that of RetinaNet model algorithm, the detection speed of SACPR model method was 5.6 times of that of RetinaNet model.
Key words: bridge    image recognition    SACPR    bolt detection    self-attention mechanism
收稿日期: 2021-03-29 出版日期: 2022-05-31
CLC:  TU 997  
基金资助: 高铁联合基金资助项目（U1934209）；中国铁路总公司系统性重大课题资助项目（P2018G002）；中国铁道科学研究院集团有限公司科研项目重大课题（2020YJ087）；安徽省引江济淮集团公司科研项目（YJJH-ZT-ZX-29210923429）
通讯作者: 赵欣欣     E-mail: juxc2008@163.com;xyzxx000@163.com
作者简介: 鞠晓臣（1982—），男，副研究员，从事桥梁运维管养研究. orcid.org/0000-0001-8460-6816. E-mail: juxc2008@163.com
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　　
鞠晓臣
赵欣欣
钱胜胜

引用本文:

鞠晓臣,赵欣欣,钱胜胜. 基于自注意力机制的桥梁螺栓检测算法[J]. 浙江大学学报(工学版), 2022, 56(5): 901-908.

Xiao-chen JU,Xin-xin ZHAO,Sheng-sheng QIAN. Self-attention mechanism based bridge bolt detection algorithm. Journal of ZheJiang University (Engineering Science), 2022, 56(5): 901-908.

链接本文:

https://www.zjujournals.com/eng/CN/10.3785/j.issn.1008-973X.2022.05.007    或    https://www.zjujournals.com/eng/CN/Y2022/V56/I5/901

图 1  螺栓检测数据集图像样本示例
图 2  SACPR算法框架示意图
图 3  螺栓检测数据集数据预处理
IR/像素 AS EF NOFC NOROS CSS
$ 608\times 608\times 3 $ 卷积 ? 32 1 2
$ 304\times 304\times 32 $ bottleneck 1 16 1 1
$ 304\times 304\times 16 $ bottleneck 6 24 2 2
$ 152\times 152\times 24 $ bottleneck 6 32 3 1
表 1  SACPR算法中的主干网络架构参数
图 4  螺栓位置预测示意图
算法 mAP/ % v/(帧·s?1)
YOLOv3[19]
（Backbone: darknet19） 		80.56 25.00
YOLOv3[19]
(Backbone: MobileNetV2) 		78.81 36.00
Faster-RCNN[18] 87.71 5.80
RetinaNet[17] 93.89 7.94
SACPR 93.91 45.00
表 2  不同算法的对比实验结果
图 5  不同模型的螺栓图片可视化检测结果比较
图 6  难例样本检测的螺栓可视化结果对比
1 张清华, 卜一之, 李乔 正交异性钢桥面板疲劳问题的研究进展[J]. 中国公路学报, 2017, 30 (3): 14- 30
ZHANG Qing-hua, BU Yi-zhi, LI Qiao Review on fatigue problems of orthotropic steel bridge deck[J]. China Journal of Highway and Transport, 2017, 30 (3): 14- 30
doi: 10.3969/j.issn.1001-7372.2017.03.002
2 鞠晓臣 U肋-面板全熔透焊接接头疲劳性能试验研究[J]. 中国公路学报, 2019, 32 (11): 176- 182
JU Xiao-chen Experimental study on fatigue performance of full penetration welded joints of Ushaped rib-plate[J]. China Journal of Highway and Transport, 2019, 32 (11): 176- 182
3 赵欣欣, 刘晓光, 潘永杰 正交异性钢桥面板纵肋腹板与面板连接构造的疲劳试验研究[J]. 中国铁道科学, 2013, 32 (2): 41- 45
ZHAO Xin-xin, LIU Xiao-guang, PAN Yong-jie Fatigue test study on the joint structure between the deck and longitudinal rib web of orthotropic steel bridge deck[J]. China Railway Science, 2013, 32 (2): 41- 45
4 钟新谷, 彭雄, 沈明燕 基于无人飞机成像的桥梁裂缝宽度识别可行性研究[J]. 土木工程学报, 2019, 52 (4): 52- 61
ZHONG Xin-gu, PENG Xiong, SHEN Ming-yan Feasibility study on bridge crack width identification based on UAV imaging[J]. China Civil Engineering Journal, 2019, 52 (4): 52- 61
5 CHOU J C, O’NEILLWENDE A, CHENGH D Pavement distress evaluation using fuzzy logic and moment in variants[J]. Transportation Research Record, 1995, 1505 (1): 39- 46
6 NGUYEN N H, KAM T Y, CHENG P Y An automatic approach for accurate edge detection of concrete crack utilizing 2D geometric features of crack[J]. Journal of Signal Processing System, 2014, 77 (3): 221- 240
doi: 10.1007/s11265-013-0813-8
7 CHA Y J, CHOI W, BÜYÜKÖZTÜRK O Deep learning-based crack damage detection using convolutional neural networks[J]. Compute-aided Civil and Infrastructure Engineering, 2017, 32 (5): 361- 378
doi: 10.1111/mice.12263
8 沙爱民, 童峥, 高杰 基于卷积神经网络的路表病害识别与测量[J]. 中国公路学报, 2018, 31 (1): 1- 10
SHA Ai-min, TONG Zheng, GAO Jie Road surface disease recognition and measurement based on convolutional neural network[J]. China Journal of Highway and Transport, 2018, 31 (1): 1- 10
doi: 10.3969/j.issn.1001-7372.2018.01.001
9 韩晓健, 赵志成, 沈泽江 卷积神经网络在桥梁结构表面病害检测中的应用研究[J]. 结构工程师, 2019, 35 (2): 106- 111
HAN Xiao-jian, ZHAO Zhi-cheng, SHEN Ze-jiang Research on the application of convolutional neural network in the detection of bridge structure surface diseases[J]. Structural Engineers, 2019, 35 (2): 106- 111
doi: 10.3969/j.issn.1005-0159.2019.02.014
10 赵欣欣, 钱胜胜, 刘晓光 基于卷积神经网络的铁路桥梁高强螺栓缺失图像识别方法[J]. 中国铁道科学, 2018, 39 (4): 56- 62
ZHAO Xin-xin, QIAN Sheng-sheng, LIU Xiao-guang Recognition method for missing images of railway bridge high-strength bolts based on convolutional neural network[J]. China Railway Science, 2018, 39 (4): 56- 62
doi: 10.3969/j.issn.1001-4632.2018.04.09
11 李良福, 马卫飞, 李丽 基于深度学习的桥梁裂缝检测算法研究[J]. 自动化学报, 2019, 6: 1- 16
LI Liang-fu, MA Wei-fei, LI Li Research on bridge crack detection algorithm based on deep learning[J]. Acta Automatica Sinica, 2019, 6: 1- 16
12 CHA Y J, CHOI W, SUH G Autonomous structural visual inspection using region-based deep learning for detecting multiple damage types[J]. Computer-aided Civil and Infrastructure Engineering, 2018, 33 (4): 1- 17
13 ZHANG C, CHANG C, JAMSHIDI M Concrete bridge surface damage detection using a single-stage detector[J]. Computer-aided Civil and Infrastructure Engineering, 2020, 35 (4): 389- 409
14 LUVIZON D C, PICARD D, TABIA H HediTabia: multi-task deep learning for real-time 3D human pose estimation and action recognition[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2021, 43 (8): 2752- 2764
15 ZHOU X, WANG D, KRÄHENBÜHL P. Objects as points [EB/OL]. [2019-04-16]. https://arxiv.org/abs/1904.07850.
16 SANDLER M, HOWARD A, ZHU M. Mobilenetv2: invert residuals and linear bottlenecks [C]// Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. Salt Lake City: IEEE, 2018, 4510-4520.
17 LIN T Y, GOYAL P, GIRSHICK R, et al. Focal loss for dense object detection [C]// Proceedings of the IEEE International Conference on Computer Vision. Venice: IEEE, 2017: 2980-2988.
18 REN S, HE K, 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, 2016, 39 (6): 1137- 1149
[1] 汪劲丰,杨松伟,亢阳阳,向华伟. 分阶段施工中钢箱梁制造参数的通用计算方法[J]. 浙江大学学报(工学版), 2022, 56(3): 550-557.
[2] 刘英莉,吴瑞刚,么长慧,沈韬. 铝硅合金实体关系抽取数据集的构建方法[J]. 浙江大学学报(工学版), 2022, 56(2): 245-253.
[3] 向胜涛,王达. 基于改进量子遗传算法的模型交互修正方法[J]. 浙江大学学报(工学版), 2022, 56(1): 100-110.
[4] 冀伟,邵天彦. 多跨连续梁桥顶推施工双导梁的优化分析[J]. 浙江大学学报(工学版), 2021, 55(7): 1289-1298.
[5] 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.
[6] 谢旭,黄文彤,冀龙飞,王天佳. 断层破裂过程对减隔震桥梁地震反应的影响[J]. 浙江大学学报(工学版), 2021, 55(12): 2225-2233.
[7] 陈巧红,陈翊,李文书,贾宇波. 多尺度SE-Xception服装图像分类[J]. 浙江大学学报(工学版), 2020, 54(9): 1727-1735.
[8] 廖小伟,王元清,吴剑国,石永久. 低温环境下十字形非传力角焊缝接头的疲劳性能[J]. 浙江大学学报(工学版), 2020, 54(10): 2018-2026.
[9] 王立国,邵旭东,曹君辉,陈玉宝,何广,王洋. 基于超短栓钉的钢-超薄UHPC组合桥面性能[J]. 浙江大学学报(工学版), 2020, 54(10): 2027-2037.
[10] 戴显荣,王路,王昌将,王晓阳,沈锐利. 多塔悬索桥全竖向摩擦板式抗滑方案[J]. 浙江大学学报(工学版), 2019, 53(9): 1697-1703.
[11] 李明, 刘扬, 杨兴胜. 考虑轴重相关的随机车流荷载效应[J]. 浙江大学学报(工学版), 2019, 53(1): 78-88.
[12] 赵人达, 贾毅, 占玉林, 王永宝, 廖平, 李福海, 庞立果. 强震区多跨长联连续梁桥减隔震设计[J]. 浙江大学学报(工学版), 2018, 52(5): 886-895.
[13] 郐艳荣, 齐梅兰, 李金钊. 近海岸桥梁下部结构波浪力分析[J]. 浙江大学学报(工学版), 2018, 52(12): 2356-2364.
[14] 许翔, 黄侨, 任远. 局部变权和云理论在悬索桥综合评估中的应用[J]. 浙江大学学报(工学版), 2017, 51(8): 1544-1550.
[15] 汪炳, 黄侨, 邹韵, 李文贤. 开孔板连接件剩余承载力计算模型及试验验证[J]. 浙江大学学报(工学版), 2017, 51(8): 1537-1543.