Please wait a minute...
Shopping Cart Email List Chinese
Advanced Search Fig/Tab
Frontiers of Structural and Civil Engineering

ISSN 2095-2430

ISSN 2095-2449(Online)

CN 10-1023/X

Postal Subscription Code 80-968

2018 Impact Factor: 1.272

Featured Articles  |  Most Downloaded  |  Most Reading
Online Submission Subscribe to Our RSS Content Feed
Front. Struct. Civ. Eng.    2022, Vol. 16 Issue (5) : 564-575    https://doi.org/10.1007/s11709-022-0829-x
RESEARCH ARTICLE
Efficient Identification of water conveyance tunnels siltation based on ensemble deep learning
Xinbin WU1(), Junjie LI1,2, Linlin WANG1
1. Faculty of Infrastructure Engineering, Dalian University of Technology, Dalian 116024, China
2. College of Water Conservancy and Hydropower Engineering, Hohai University, Nanjing 210098, China
 Download: PDF(2421 KB)   HTML
 Export: BibTeX | EndNote | Reference Manager | ProCite | RefWorks
Abstract

The inspection of water conveyance tunnels plays an important role in water diversion projects. Siltation is an essential factor threatening the safety of water conveyance tunnels. Accurate and efficient identification of such siltation can reduce risks and enhance safety and reliability of these projects. The remotely operated vehicle (ROV) can detect such siltation. However, it needs to improve its intelligent recognition of image data it obtains. This paper introduces the idea of ensemble deep learning. Based on the VGG16 network, a compact convolutional neural network (CNN) is designed as a primary learner, called Silt-net, which is used to identify the siltation images. At the same time, the fully-connected network is applied as the meta-learner, and stacking ensemble learning is combined with the outputs of the primary classifiers to obtain satisfactory classification results. Finally, several evaluation metrics are used to measure the performance of the proposed method. The experimental results on the siltation dataset show that the classification accuracy of the proposed method reaches 97.2%, which is far better than the accuracy of other classifiers. Furthermore, the proposed method can weigh the accuracy and model complexity on a platform with limited computing resources.
Keywords water conveyance tunnels      siltation images      remotely operated vehicles      deep learning      ensemble learning      computer vision     
Corresponding Author(s): Xinbin WU   
About author:

Tongcan Cui and Yizhe Hou contributed equally to this work.
Just Accepted Date: 28 April 2022   Online First Date: 02 August 2022    Issue Date: 30 August 2022
 Cite this article:   
Xinbin WU,Junjie LI,Linlin WANG. Efficient Identification of water conveyance tunnels siltation based on ensemble deep learning[J]. Front. Struct. Civ. Eng., 2022, 16(5): 564-575.
 URL:  
https://academic.hep.com.cn/fsce/EN/10.1007/s11709-022-0829-x
https://academic.hep.com.cn/fsce/EN/Y2022/V16/I5/564
Fig.1  Image acquisition system: (a) underwater inspection system; (b) ground control system.
Fig.2  An overview of ROV system.
Fig.3  Different types of samples from different perspectives: (a) severe siltation; (b) general siltation; (c) no siltation and (d) no targets.
dataset category number
train5000 severe siltation 1250
general siltation 1250
no siltation 1250
no targets 1250
test1000 severe siltation 250
general siltation 250
no siltation 250
no targets 250
Tab.1  Dataset for training and testing
Fig.4  An overview of proposed Silt-nets stacking framework.
layer size stride operator output shape parameters
2242 × 3 s1 2*(conv2d, 3 × 3) 2242 × 64 38720
2242 × 64 s2 max pooling2d 1122 × 64 –
1122 × 64 s1 2*(conv2d, 3 × 3) 1122 × 128 221440
1122 × 128 s2 max pooling2d 562 × 128 –
562 × 128 s1 3*(conv2d, 3 × 3) 562 × 256 1475328
562 × 256 s2 max pooling2d 282 × 256 –
282 × 256 s1 3*(conv2d, 3 × 3) 282 × 512 5899776
282 × 512 s2 max pooling2d 142 × 512 –
142 × 512 s1 3*(conv2d, 3 × 3) 142 × 512 7079424
142 × 512 s2 max pooling2d 72 × 512 –
72 × 512 – flatten 25088 –
25088 – 2*dense 4096 119545856
4096 – dense 4 16388
Tab.2  VGG16 architecture
layer size stride operator output shape parameters
2242 × 3 s1 conv2d, 3 × 3 2242 × 16 512
2242 × 16 s2 max pooling2d 1122 × 16 –
1122 × 16 s1 2*(conv2d, 3 × 3) 1122 × 32 14144
1122 × 32 s2 max pooling2d 562 × 32 –
562 × 32 s1 2*(conv2d, 3 × 3) 562 × 64 55936
562 × 64 s2 max pooling2d 282 × 64 –
282 × 64 – flatten 43264 –
43264 – dense 512 22151680
512 – dense 4 2052
Tab.3  Silt-net architecture
stage variable hyper-parameters
stage 1 RMSprop Lr = 0.001
epoch 100
batch size 32
stage 2 Adam Lr = 0.001
L2 regularization 0.0001
early stopping patience = 20
dropout ratio = 0.5
Tab.4  Experimental hyper-parameter setting
Fig.5  Loss curves of base classifiers and meta-classifier in training and validation process: (a) base classifier 1; (b) base classifier 2; (c) base classifier 3; (d) meta-classifier.
Fig.6  Classification results of different classifiers.
method RBF-SVM Silt-net VGG16 our proposed
heavy silt 94.8% 97.2% 94.0% 98.0%
general silt 96.0% 96.8% 98.8% 99.2%
no silt 84.4% 96.0% 96.4% 95.6%
no targets 84.0% 90.4% 96.8% 94.8%
OA 89.8% 95.1% 96.5% 96.9%
Tab.5  Classification performance of each classifier in various categories
method Silt-net VGG16 our proposed
params (M) 22.22 134.28 75.1
FLOPs (G) 0.4 15.5 1.2
Tab.6  Comparison of model complexity between different models
item base learner 1 base learner 2 base learner 3 DIV ensemble results
1 94.3% 94.6% 95.1% 0.042 96.4%
2 95.1% 95.8% 96.0% 0.030 96.8%
3 94.3% 94.6% 96.3% 0.043 96.9%
4 94.3% 95.1% 96.3% 0.040 96.9%
5 95.8% 96.0% 96.3% 0.032 97.2%
Tab.7  Ensemble results of different base learner combinations
Fig.7  Results of different ensemble methods under different numbers of base learners.
1 X M Zhu, T H Wang, Y B Liu, T Huang. A new defect detection technology for long-distance water conveyance tunnel. Water Resources and Hydropower Engineering, 2010, 41( 12): 78– 81
2 R Moughamian M McLeod. Pardee tunnel inspection and condition assessment. In: Conference on Pipeline Engineering––Concepts in Harmony (PIPELINES). Nashville: American Society of Civil Engineers, 2019
3 V A M Jorge P D D Gava J R B F Silva T M Mancilha W Vieira G J Adabo C L Nascimento. VITA1: An unmanned underwater vehicle prototype for operation in underwater tunnels. In: the 15th Annual IEEE International Systems Conference (Syscon 2021). IEEE, 2021
4 J T Lai. Research and application of underwater full coverage unmanned detection technology for large diameter and long diversion tunnel. Yangtze River, 2020, 51( 05): 228– 232
5 Y Pan, L M Zhang, X G Wu, M J Skibniewski. Multi-classifier information fusion in risk analysis. Information Fusion, 2020, 60 : 121– 136
https://doi.org/10.1016/j.inffus.2020.02.003
6 H Pan, M Azimi, F Yan, Z Lin. Time-frequency-based data-driven structural diagnosis and damage detection for cable-stayed bridges. Journal of Bridge Engineering, 2018, 23( 6): 04018033
https://doi.org/10.1061/(ASCE)BE.1943-5592.0001199
7 S F Wirtz, N Beganovic, D Soffker. Investigation of damage detectability in composites using frequency-based classification of acoustic emission measurements. Structural Health Monitoring, 2019, 18( 4): 1207– 1218
https://doi.org/10.1177/1475921718791894
8 H Babajanian Bisheh, G Ghodrati Amiri, M Nekooei, E Darvishan. Damage detection of a cable-stayed bridge using feature extraction and selection methods. Structure and Infrastructure Engineering, 2019, 15( 9): 1165– 1177
https://doi.org/10.1080/15732479.2019.1599964
9 B Kurian R Liyanapathirana. Machine learning techniques for structural health monitoring. In: the 13th International Conference on Damage Assessment of Structures. Porto: Springer, 2020
10 N Mechbal, J S Uribe, M Rebillat. A probabilistic multi-class classifier for structural health monitoring. Mechanical Systems and Signal Processing, 2015, 60-61 : 106– 123
https://doi.org/10.1016/j.ymssp.2015.01.017
11 Y Huang, J L Beck, H Li. Hierarchical sparse Bayesian learning for structural damage detection: Theory, computation and application. Structural Safety, 2017, 64 : 37– 53
https://doi.org/10.1016/j.strusafe.2016.09.001
12 M Azimi, A D Eslamlou, G Pekcan. Data-driven structural health monitoring and damage detection through deep learning: State-of-the-art review. Sensors (Basel), 2020, 20( 10): 2778
https://doi.org/10.3390/s20102778
13 J Devlin M W Chang K Lee K Toutanova. BERT: Pre-training of deep bidirectional transformers for language understanding, 2018, arXiv:1810.04805
14 A Bochkovskiy C Y Wang H Y M J Liao. YOLOv4: Optimal speed and accuracy of object detection. 2020, arXiv:2004.10934
15 L C E Chen, Y K Zhu, G Papandreou, F Schroff, H Adam. Encoder-decoder with atrous separable convolution for semantic image segmentation. In: the 15th european conference on computer vision (ECCV). Munich: Springer, 2018, 833– 851
16 N Moritz T Hori J Le Roux. Triggered attention for end-to-end speech recognition, In: 2019 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP). Brighton: IEEE, 2019
17 C Anitescu, E Atroshchenko, N Alajlan, T Rabczuk. Artificial neural network methods for the solution of second order boundary value problems. Computers, Materials and Continua, 2019, 59( 1): 345– 359
18 H W Guo, X Y Zhuang, T Rabczuk. A deep collocation method for the bending analysis of kirchhoff plate. Computers, Materials and Continua, 2019, 59( 2): 433– 456
19 E Samaniego, C Anitescu, S Goswami, V M Nguyen-Thanh, H Guo, K Hamdia, X Zhuang, T Rabczuk. An energy approach to the solution of partial differential equations in computational mechanics via machine learning: Concepts, implementation and applications. Computer Methods in Applied Mechanics and Engineering, 2020, 362 : 112790
20 X W Ye, T Jin, C B Yun. A review on deep learning-based structural health monitoring of civil infrastructures. Smart Structures and Systems, 2019, 24( 5): 567– 585
21 L X Luo, M Q Feng, J P Wu, R Y Leung. Autonomous pothole detection using deep region-based convolutional neural network with cloud computing. Smart Structures and Systems, 2019, 24( 6): 745– 757
22 Y J Cha, W Choi, G Suh, S Mahmoudkhani, O Büyüköztürk. Autonomous structural visual inspection using region-based deep learning for detecting multiple damage types. Computer-Aided Civil and Infrastructure Engineering, 2018, 33( 9): 731– 747
https://doi.org/10.1111/mice.12334
23 P F Shi, X N Fan, J J Ni, G Wang. A detection and classification approach for underwater dam cracks. Structural Health Monitoring, 2016, 15( 5): 541– 554
https://doi.org/10.1177/1475921716651039
24 X Liang. Image-based post-disaster inspection of reinforced concrete bridge systems using deep learning with Bayesian optimization. Computer-Aided Civil and Infrastructure Engineering, 2019, 34( 5): 415– 430
https://doi.org/10.1111/mice.12425
25 P Savino, F Tondolo. Automated classification of civil structure defects based on convolutional neural network. Frontiers of Structural and Civil Engineering, 2021, 15( 2): 305– 317
https://doi.org/10.1007/s11709-021-0725-9
26 T G Dietterich. Ensemble methods in machine learning. In: 1st International Workshop on Multiple Classifier Systems. Cagliari: Springer Science & Business Media, 2000
27 D T Bui, T C Ho, B Pradhan, B T Pham, V H Nhu, I Revhaug. GIS-based modeling of rainfall-induced landslides using data mining-based functional trees classifier with AdaBoost, Bagging, and MultiBoost ensemble frameworks. Environmental Earth Sciences, 2016, 75( 14): 1– 22
28 S Tsiapoki, O Bahrami, M W Hackell, J P Lynch, R Rolfes. Combination of damage feature decisions with adaptive boosting for improving the detection performance of a structural health monitoring framework: Validation on an operating wind turbine. Structural Health Monitoring, 2021, 20( 2): 637– 660
https://doi.org/10.1177/1475921720909379
29 P R Kadavi, C W Lee, S Lee. Application of ensemble-based machine learning models to landslide susceptibility mapping. Remote Sensing, 2018, 10( 8): 1252
https://doi.org/10.3390/rs10081252
30 Z R Li J Q Guo W S Liang X Xie G Zhang S Wang. Structural health monitoring based on realadaboost algorithm in wireless sensor networks. In: the 9th International Conference on Wireless Algorithms, Systems, and Applications (WASA). Harbin: Springer, 2014
31 R D Christ R Wernli. The ROV Manual: A User Guide for Observation Class Remotely Operated Vehicles. Oxford: Elsevier, 2013
32 M Vukšić, S Josipović, A Čorić, A Kraljević. Underwater ROV as inspection and development platform. Transactions on Maritime Science, 2017, 6( 1): 48– 54
33 Y LeCun, Y Bengio, G Hinton. Deep learning. Nature, 2015, 521( 7553): 436– 444
https://doi.org/10.1038/nature14539
34 S Dorafshan, R J Thomas, M Maguire. Comparison of deep convolutional neural networks and edge detectors for image-based crack detection in concrete. Construction & Building Materials, 2018, 186 : 1031– 1045
https://doi.org/10.1016/j.conbuildmat.2018.08.011
35 C A Perez-Ramirez, J P Amezquita-Sanchez, M Valtierra-Rodriguez, H Adeli, A Dominguez-Gonzalez, R J Romero-Troncoso. Recurrent neural network model with Bayesian training and mutual information for response prediction of large buildings. Engineering Structures, 2019, 178 : 603– 615
https://doi.org/10.1016/j.engstruct.2018.10.065
36 C S N Pathirage, J Li, L Li, H Hao, W Liu, P Ni. Structural damage identification based on autoencoder neural networks and deep learning. Engineering Structures, 2018, 172 : 13– 28
https://doi.org/10.1016/j.engstruct.2018.05.109
37 X Zhuang, H Guo, N Alajlan, H Zhu, T Rabczuk. Deep autoencoder based energy method for the bending, vibration, and buckling analysis of Kirchhoff plates with transfer learning. European Journal of Mechanics—A/Solids, 2021, 87 : 104225
38 W Zhang, X Li, X D Jia, H Ma, Z Luo, X Li. Machinery fault diagnosis with imbalanced data using deep generative adversarial networks. Measurement, 2020, 152 : 107377
https://doi.org/10.1016/j.measurement.2019.107377
39 P Tamilselvan Y B Wang P F Wang. Deep belief network based state classification for structural health diagnosis. In: IEEE Aerospace Conference, Big Sky. Montana: IEEE, 2012
40 K Simonyan A J C S Zisserman. Very deep convolutional networks for large-scale image recognition. In: International Conference on Learning Representations. San Diego: OpenReview, 2015
41 Z H Zhou. Ensemble Methods: Foundations and Algorithms. Boca Raton: Taylor & Francis, 2012
42 P Molchanov S Tyree T Karras T Aila J Kautz. Pruning convolutional neural networks for resource efficient inference. 2016, arXiv:1611.06440
43 Y X Bi. The impact of diversity on the accuracy of evidential classifier ensembles. International Journal of Approximate Reasoning, 2012, 53( 4): 584– 607
https://doi.org/10.1016/j.ijar.2011.12.011
[1] Shuai TENG, Gongfa CHEN, Shaodi WANG, Jiqiao ZHANG, Xiaoli SUN. Digital image correlation-based structural state detection through deep learning[J]. Front. Struct. Civ. Eng., 2022, 16(1): 45-56.
[2] Trong NGHIA-NGUYEN, Mamoru KIKUMOTO, Samir KHATIR, Salisa CHAIYAPUT, H. NGUYEN-XUAN, Thanh CUONG-LE. Data-driven approach to solve vertical drain under time-dependent loading[J]. Front. Struct. Civ. Eng., 2021, 15(3): 696-711.
[3] Alipujiang JIERULA, Tae-Min OH, Shuhong WANG, Joon-Hyun LEE, Hyunwoo KIM, Jong-Won LEE. Detection of damage locations and damage steps in pile foundations using acoustic emissions with deep learning technology[J]. Front. Struct. Civ. Eng., 2021, 15(2): 318-332.
[4] Harun TANYILDIZI, Abdulkadir ŞENGÜR, Yaman AKBULUT, Murat ŞAHİN. Deep learning model for estimating the mechanical properties of concrete containing silica fume exposed to high temperatures[J]. Front. Struct. Civ. Eng., 2020, 14(6): 1316-1330.
Viewed
Full text


Abstract

Cited
  Shared   
  Discussed