基于卷积神经网络和投票机制的轨道板裂缝检测

doi:10.3969/j.issn.0255-8297.2021.04.010

摘要/Abstract

摘要： 现有的检测方法对轨道板细微裂缝和夜间拍摄的裂缝图像存在误检和漏检的现象，为此提出了一种基于卷积神经网络的改进方法。将特征图分组后用注意力机制强化各组向量的特征表达，以动态聚合弱分类器预测结果的方式得到最终的裂缝置信度。借助投票机制有效降低最终的预测偏差，提升模型的鲁棒性。实验结果表明：该改进方法在减少模型参数的情况下，在裂缝数据集上的准确率提升1.6%，在CIFAR-10数据集上的准确率提升2.8%。

关键词: 裂缝检测, 卷积神经网络, 投票机制, 训练策略, 深度学习

Abstract: There are misdetections and missed detections in the crack detection of track slab or in crack pictures taken at night as using existing detection methods. For the problem, an improved method based on convolutional neural network (CNN) is proposed. In this way, high-level feature maps are divided into different groups of vectors whose feature expression would be emphasized by attention mechanism subsequently. Final confidence is accounted by aggregating the predict result of weak classifiers dynamically. With the favor of voting mechanism, predict error is reduced and robustness of model is improved effectively. Experiment results show that the improved method gains a prediction improvement of 1.6% in crack dataset and an improvement of 2.8% in CTFAR-10 dataset, in spite of the reduction of model parameters.

Key words: crack detection, convolutional neural network (CNN), voting mechanism, training tactic, deep learning

中图分类号:

TN911.73

李文举, 何茂贤, 张耀星, 陈慧玲, 李培刚. 基于卷积神经网络和投票机制的轨道板裂缝检测[J]. 应用科学学报, 2021, 39(4): 627-640.

LI Wenju, HE Maoxian, ZHANG Yaoxing, CHEN Huiling, LI Peigang. Crack Detection of Track Slab Based on Convolutional Neural Network and Voting Mechanism[J]. Journal of Applied Sciences, 2021, 39(4): 627-640.

参考文献

[1] 史云飞, 何越磊, 张志远, 等. 基于图像的CRTS II型轨道板面横向裂纹扩展特征分析[J]. 铁道科学与工程学报, 2017, 14(10):2059-2064. Shi Y F, He Y L, Zhang Z Y, et al. Characteristic analysis of transverse crack growth for CRTS II slab track based on image processing[J]. Journal of Railway Science and Engineering, 2017, 14(10):2059-2064. (in Chinese)
[2] 李培刚. CRTSII型板式轨道层间损伤及其影响研究[D]. 成都:西南交通大学, 2015.
[3] Oliveira H, Correia P L. Supervised crack detection and classification in images of road pavement flexible surfaces[M]. InTech, 2009.
[4] 钟鹏飞, 车爱兰, 冯少孔, 等. 高速铁路线下结构典型病害分析及快速无损检测方法研究[J]. 振动与冲击, 2017, 36(11):154-160. Zhong P F, Che A L, Feng S K, et al. Typical defects analysis and nondestructive detection method for undertrack structures of high speed railways[J]. Journal of Vibration and Shock, 2017, 36(11):154-160. (in Chinese)
[5] 廖红建, 朱庆女, 昝月稳, 等. 基于探地雷达的高铁无砟轨道结构层病害检测[J]. 西南交通大学学报, 2016, 51(1):8-13. Liao H J, Zhu Q N, Zan Y W, et al. Detection of ballastless track diseases in high-speed railway based on ground penetrating radar[J]. Journal of Southwest Jiaotong University, 2016, 51(1):8-13. (in Chinese)
[6] Azulay A, Weiss Y. Why do deep convolutional networks generalize so poorly to small image transformations?[J]. Journal of Machine Learning Research, 2019, 20(184):1-25.
[7] Dai J, Li Y, He K, et al. R-FCN:object detection via region-based fully convolutional networks[C]//Advances in Neural Information Processing Systems, 2016:379-387.
[8] 王丽苹, 高瑞贞, 张京军, 等. 基于卷积神经网络的混凝土路面裂缝检测[J]. 计算机科学, 2019, 46(Suppl.2):584-589. Wang L P, Gao R Z, Zhang J J, et al. Crack detection of concrete pavement based on convolutional neural network[J]. Computer Science, 46(Suppl.2):584-589. (in Chinese)
[9] 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. Chai X S, Zhu X Y, Li J C, et al. Tunnel lining crack identification algorithm based on deep convolutional neural network[J]. Railway Engineering, 2018, 58(6):60-65. (in Chinese)
[10] 李良福, 马卫飞, 李丽, 等. 基于深度学习的桥梁裂缝检测算法研究[J]. 自动化学报, 2019, 45(9):1727-1742. Li L F, Ma W F, Li L, et al. Research on detection algorithm for bridge cracks based on deep learning[J]. Acta Automatica Sinica, 2019, 45(9):1727-1742. (in Chinese)
[11] Zhou S, Song W. Deep learning-based roadway crack classification using laser-scanned range images:a comparative study on hyperparameter selection[J]. Automation in Construction, 2020, 114:103171-103185.
[12] 杜清超, 钟伟, 郑佩莹. 基于数字图像处理的轨道梁裂缝检测技术[J]. 四川建筑, 2019, 39(4):95-97. Du Q C, Zhong W, Zheng P Y. Crack detection technology of track beam based on digital image processing[J]. Sichuan Architecture, 2019, 39(4):95-97. (in Chinese)
[13] 薛峰, 赵丽科, 柴雪松, 等. 基于图像处理的铁路轨道板裂缝检测研究[J]. 铁道建筑, 2015(12):123-126. Xue F, Zhao L K, Chai X S, et al. Study on detecting crack in railway track slab based on image processing technology[J]. Railway Engineering, 2015(12):123-126. (in Chinese)
[14] 王登涛, 李再帏, 何越磊, 等. 基于热成像的高速铁路轨道板表面裂缝检测方法研究[J]. 铁道标准设计, 2020, 64(7):22-28. Wang D T, Li Z W, He Y L, et al. Research on the surface crack detection method of high-speed railway track plate based on thermal imaging[J]. Railway Standard Design, 2020, 64(7):22-28. (in Chinese)
[15] Krizhevsky A, Sutskever I, Hinton G E. ImageNet classification with deep convolutional neural networks[J]. Advances in Neural Information Processing Systems, 2012, 25:1097-1105.
[16] Simonyan K, Zisserman A. Very deep convolutional networks for large-scale image recognition[C]//IEEE Conference on Computer Vision and Pattern Recognition, 2014:178-192.
[17] Szegedy C, Liu W, Jia Y, et al. Going deeper with convolutions[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2015:1-9.
[18] He K, Zhang X, Ren S, et al. Deep residual learning for image recognition[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2016:770-778.
[19] He T, Zhang Z, Zhang H, et al. Bag of tricks for image classification with convolutional neural networks[C]//IEEE Conference on Computer Vision and Pattern Recognition, 2019:558-567.
[20] Long J, Shelhamer E, Darrell T. Fully convolutional networks for semantic segmentation[J]. IEEE Transactions on Pattern Analysis & Machine Intelligence, 2014, 39(4):640-651.
[21] Xie S, Tu Z. Holistically-nested edge detection[J]. International Journal of Computer Vision, 2017, 125(1/2/3):3-18.
[22] Liu Y, Cheng M M, Hu X, et al. Richer convolutional features for edge detection[C]//IEEE Conference on Computer Vision and Pattern Recognition, 2016:3000-3009.
[23] Lin T Y, Dollár P, Girshick R, et al. Feature pyramid networks for object detection[C]//IEEE Conference on Computer Vision and Pattern Recognition, 2017:2117-2125.
[24] Hu J, Shen L, Sun G. Squeeze-and-excitation networks[C]//IEEE Conference on Computer Vision and Pattern Recognition, 2018:7132-7141.
[25] Wang F, Jiang M, Qian C, et al. Residual attention network for image classification[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2017:3156-3164.
[26] Li X, Wang W, Hu X, et al. Selective kernel networks[C]//IEEE Conference on Computer Vision and Pattern Recognition, 2019:510-519.
[27] Goyal P, Dollár P, Girshick R, et al. Accurate, large minibatch SGD:training ImageNet in 1 hour[J]. arXiv e-prints:abs/1706.02677, 2017[2020-09-15]. https://arxiv.org/pdf/1706.02677.pdf.
[28] Loshchilov I, Hutter F. SGDR:Stochastic gradient descent with warm restarts[J]. Learning, 2016, 10:3-18.
[29] Ioffe S, Szegedy C. Batch normalization:accelerating deep network training by reducing internal covariate shift[C]//International Conference on Machine Learning, 2015:448-456.
[30] He K, Zhang X, Ren S, et al. Delving deep into rectifiers:surpassing human-level performance on ImageNet classification[C]//IEEE Conference on Computer Vision and Pattern Recognition, 2015:1026-1034.
[31] Lee C Y, Gallagher P W, Tu Z. Generalizing pooling functions in convolutional neural networks:mixed, gated, and tree[C]//IEEE Trans Pattern Anal Mach Intell. 2018, 40(4):863-875.
[32] Krizhevsky A, Hinton G. Learning multiple layers of features from tiny images[D]. Toronto:University of Toronto, 2009.
[33] Chang J R, Chen Y S. Batch-normalized Maxout network in network[J/OL]. ArXiv:abs/1511.02583. (2015-01-09)[2020-0915]. https://arxiv.org/pdf/1511.02583.pdf.
[34] Springenberg J T, Dosovitskiy A, Brox T, et al. Striving for simplicity:the all convolutional net[C]//International Conference on Learning Representations:2014:1-14.