Detection and classification of pipe defects based on pipe-extended feature pyramid network

被引:10
作者
Guo, Wenhao [1 ,2 ,3 ]
Zhang, Xing [2 ,3 ]
Zhang, Dejin [2 ,3 ]
Chen, Zhipeng [2 ,3 ]
Zhou, Baoding [3 ,4 ]
Huang, Dingfa [1 ]
Li, Qingquan [1 ,2 ,3 ]
机构
[1] Southwest Jiaotong Univ, Fac Geosci & Environm Engn, Chengdu 611756, Sichuan, Peoples R China
[2] Shenzhen Univ, Sch Architecture & Urban Planning, Shenzhen 518060, Guangdong, Peoples R China
[3] Shenzhen Univ, Guangdong Key Lab Urban Informat, Shenzhen 518060, Peoples R China
[4] Shenzhen Univ, Inst Urban Smart Transportat & Safty Maintenance, Shenzhen 518060, Peoples R China
基金
中国国家自然科学基金;
关键词
Pipe defects; Underground infrastructure; Textural features; Defects detection; Defects classification; SEWER; MODEL;
D O I
10.1016/j.autcon.2022.104399
中图分类号
TU [建筑科学];
学科分类号
0813 ;
摘要
In image-based pipe defect detection research, the effective utilization of the information in the two-dimension (2D) image is directly related to the sampling of the image. The existing inspection methods do not analyze the pipeline imaging but rather directly use the object detection method for defect detection, resulting in a bottleneck problem for the accuracy. In this study, the pipeline imaging was analyzed. It was found that effective sampling of the defect texture within the edge region of the image could improve defect detection accuracy. An image sampling framework, pipe-extended feature pyramid network (P-EFPN), was constructed, and the super-resolution (SR) module was embedded for texture extraction to obtain rich defect texture information and provide image sampling support for pipe defect detection. The defect dataset contains deformation, corrosion, and crack. In the faster region-convolutional neural network (R-CNN) model with Resnet-101 as the backbone, the mean average precision (mAP) of the P-EFPN model was improved by 8.64% compared to the state-of-the-art feature pyramid network (FPN) model. The proposed method improves the accuracy of defect detection by capturing more textures in the edge regions of the image. Compared with existing image sampling methods, the proposed sampling method is more suitable for pipe defect detection.
引用
收藏
页数:10
相关论文
共 44 条
  • [11] Automated defect detection tool for closed circuit television (cctv) inspected sewer pipelines
    Hawari, Alaa
    Alamin, Mazen
    Alkadour, Firas
    Elmasry, Mohamed
    Zayed, Tarek
    [J]. AUTOMATION IN CONSTRUCTION, 2018, 89 : 99 - 109
  • [12] Jun Z, 2013, P INT C PIPELINES TR, P1942, DOI [10.1061/9780784412619.195, DOI 10.1061/9780784412619.195]
  • [13] Khalifa I., 2014, INT J COMPUT APPL, V87, P25, DOI DOI 10.5120/15238-3779
  • [14] Kim KI, 2002, LECT NOTES COMPUT SC, V2388, P293
  • [15] Panoptic Feature Pyramid Networks
    Kirillov, Alexander
    Girshick, Ross
    He, Kaiming
    Dollar, Piotr
    [J]. 2019 IEEE/CVF CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION (CVPR 2019), 2019, : 6392 - 6401
  • [16] Innovative method for assessment of underground sewer pipe condition
    Koo, Dae-Hyun
    Ariaratnam, Samuel T.
    [J]. AUTOMATION IN CONSTRUCTION, 2006, 15 (04) : 479 - 488
  • [17] Automated defect classification in sewer closed circuit television inspections using deep convolutional neural networks
    Kumar, Srinath S.
    Abraham, Dulcy M.
    Jahanshahi, Mohammad R.
    Iseley, Tom
    Starr, Justin
    [J]. AUTOMATION IN CONSTRUCTION, 2018, 91 : 273 - 283
  • [18] MsCNN: A Deep Learning Framework for P300-Based Brain-Computer Interface Speller
    Kundu, Sourav
    Ari, Samit
    [J]. IEEE TRANSACTIONS ON MEDICAL ROBOTICS AND BIONICS, 2020, 2 (01): : 86 - 93
  • [19] Two-dimensional materials in functional three-dimensional architectures with applications in photodetection and imaging
    Lee, Wonho
    Liu, Yuan
    Lee, Yongjun
    Sharma, Bhupendra K.
    Shinde, Sachin M.
    Kim, Seong Dae
    Nan, Kewang
    Yan, Zheng
    Han, Mengdi
    Huang, Yonggang
    Zhang, Yihui
    Ahn, Jong-Hyun
    Rogers, John A.
    [J]. NATURE COMMUNICATIONS, 2018, 9
  • [20] Sewer damage detection from imbalanced CCTV inspection data using deep convolutional neural networks with hierarchical classification
    Li, Duanshun
    Cong, Anran
    Guo, Shuai
    [J]. AUTOMATION IN CONSTRUCTION, 2019, 101 : 199 - 208