Water leakage of tunnel is a common disease in all kinds of existing tunnels. Automatic, timely, and accurate detection of water leakages is of great significance to the safe operation and maintenance for all kinds of tunnels. Due to the complicated background of tunnel surfaces, it is difficult for traditional tunnel defect detection algorithms to detect such defects quickly and accurately. To address these problems, this article presents an improved BlendMask image segmentation model with top-down and bottom-up architecture which accurately extract tunnel’s water leakage area from tunnel lining images. The proposed method is validated by an experimental study, and the results are compared with those obtained by the tunnel defect segmentation methods including Mask R-CNN and DeepCrack and other deep learning methods including CondInst, SoloV2, fully convolutional network and UNet. The Recall, Precision, F1-score, Dice and mean intersection over union (mIoU) for the proposed method are superior than those by the other methods above with respect to on test images. The subjective analysis on predicted results by different methods also shows the presented method can perform well.
YuanYJiangXLiuX. Predictive maintenance of shield tunnels. Tunnelling Underground Space Tech2013; 38: 69–86.
2.
HuangH-wLiQ-tZhangD-m. Deep learning based image recognition for crack and leakage defects of metro shield tunnel. Tunnelling Underground Space Tech2018; 77: 166–176.
3.
HuCZhuHLiX. Detection of tunnel water leakage based on image processing. In: Proceedings of International conference on information technology in geo-engineering, Shanghai, China, 15 September 2010, pp. 254–262.
4.
DawoodTZhuZHZayedT. Computer vision-based model for moisture marks detection and recognition in subway networks. J Comput Civil Eng2018; 32(2): 4017079–4017081.
5.
ChengXHuXTanK, et al. Automatic detection of shield tunnel leakages based on terrestrial mobile LiDAR intensity images using deep learning. IEEE Access2021; 9: 55300–55310.
ZhaoSZhangDMHuangHW. Deep learning-based image instance segmentation for moisture marks of shield tunnel lining. Tunnelling Underground Space Tech2020; 95: 103156.
8.
ByunghyunKSoojinC. Image-based concrete crack assessment using mask and region-based convolutional neural network. Struct Control Health Monit2019; 26(8): e2381.1–e2381.15.
9.
YadongXXinyuanCMahdiS, et al. Deep learning-based automatic recognition of water leakage area in shield tunnel lining. Tunn Undergr Sp Tech2020; 104: 103524.
10.
GaoXJianMHuM, et al. Faster multi-defect detection system in shield tunnel using combination of FCN and faster RCNN. Adv Struct Eng2019; 22(13): 2907–2921.
11.
RossGJeffDTrevorD, et al. Rich feature hierarchies for accurate object detection and semantic segmentation. In: Proceedings of the 2014 IEEE conference on computer vision and pattern recognition (CVPR 2014), Columbus, OH, 23–28, June 2014, pp. 580–587. Washington, DC: IEEE Computer Society.
12.
ChenHSunKTianZ, et al. BlendMask: Top-down meets bottom-up for instance segmentation. In: Proceedings of the 2020 IEEE Conference on Computer Vision and Pattern Recognition (CVPR 2020), Seattle, WA, 14–20 June 2020. Washington, DC: IEEE Computer Society.
13.
TianZShenCChenH, et al. FCOS: fully convolutional one-stage object detection. In: Proceedings of 2019 IEEE/CVF International Conference on Computer (ICCV 2019), Seoul, South Korea, 20–26 October 2019, pp. 9626–9635. Piscataway, NJ: Institute of Electrical and Electronics Engineers Inc.
14.
DanielBChongZFanyiX, et al. YOLACT: real-time instance segmentation. In: Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV 2019), Seoul, South Korea, 20–26 October 2019. Piscataway, NJ: IEEE. Piscataway, NJ: IEEE
15.
FeIWJiangMChenQ, et al. Residual attention network for image classification. In: Proceedings of 30th IEEE Conference on Computer Vision and Pattern Recognition (CVPR 2017), Honolulu, HI, 21–26 July 2017, pp. 6450–6458. Washington, DC: IEEE Computer Society.
16.
FuJLiuJTianH, et al. Dual attention network for scene segmentation. In: Proceedings of 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR 2019), Long Beach, CA, 15–20 June 2019, pp. 3141–3149. Washington, DC: IEEE Computer Society.
17.
JieHLiSGangS, et al. Squeeze-and-excitation networks. IEEE T Pattern Anal2020; 42(8): 2011–2023.
18.
WooSParkJLeeJ-Y, et al. CBAM: convolutional block attention module. Computer Vis - ECCV 20182018; 11211: 3–19.
19.
LiuTZTengGFYuanYC, et al. Winter jujube fruit recognition method based on improved YOLO v3 under natural scene. Trans Chin Soc Agric Machinery2021; 52(5): 17–25.
20.
WangQLiuQLuoG, et al. Automated segmentation and diagnosis of pneumothorax on chest X-rays with fully convolutional multi-scale ScSE-DenseNet: a retrospective study. BMC Medical Informatics Decision Making2020; 20(Suppl 14): 317.
21.
RoyAGNavabNWachingerC. Recalibrating fully convolutional networks with spatial and channel "Squeeze and excitation" blocks. IEEE Trans Med Imaging2019; 38(2): 540–549.
22.
ChenYZhangXChenW, et al. Research on recognition of fly species based on improved RetinaNet and CBAM. IEEE Access2020; 8: 102907–102919.
23.
SuiHGHuangLHLiuCX. Detecting building façade damage caused by Earthquake using CBAM-improved mask R-CNN. Geomatics Inf Sci Wuhan Univ2020; 45(11): 1660–1668.
24.
RussellB. LabeMe: a database and web-based tool for image annotation. INT J Comput Vis2008; 77(1): 157–173.
25.
TianZShenCChenHConditional convolutions for instance segmentation. Computer Vis - ECCV 20202020; 12346: 282–298.
26.
WangXLZhangRFKongT, et al. SOLOv2: dynamic and fast instance segmentation. In: Proceedings of 34th Conference on Neural Information Processing Systems, 11–13 December 2020. NeurIPS.
27.
QinZZhengZLiQQ, et al. DeepCrack: learning hierarchical convolutional features for crack detection. IEEE T Image Process2019; 28(3): 1498–1512.
