Crack Identification for Bridge Condition Monitoring Combining Graph Attention Networks and Convolutional Neural Networks

Orthotropic steel box girders and steel bridge decks are commonly applied to bridges. Because of the coupling of original defects and alternating forces, fatigue cracks are likely to appear in the structures. In order to ensure the life span of bridges, methods for automatic crack identification are...

Full description

Saved in:
Bibliographic Details
Main Authors: Feiyu Chen, Tong Tong, Jiadong Hua, Chun Cui
Format: Article
Language:English
Published: MDPI AG 2025-05-01
Series:Applied Sciences
Subjects:
Online Access:https://www.mdpi.com/2076-3417/15/10/5452
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850127270713229312
author Feiyu Chen
Tong Tong
Jiadong Hua
Chun Cui
author_facet Feiyu Chen
Tong Tong
Jiadong Hua
Chun Cui
author_sort Feiyu Chen
collection DOAJ
description Orthotropic steel box girders and steel bridge decks are commonly applied to bridges. Because of the coupling of original defects and alternating forces, fatigue cracks are likely to appear in the structures. In order to ensure the life span of bridges, methods for automatic crack identification are needed. In this paper, we present a novel approach for crack detection and bridge condition monitoring by integrating convolutional neural networks (CNNs) with graph attention networks (GATs). At first, the original large-sized images are divided into small-sized patches, and these patches are input into a CNN architecture to extract features by decreasing dimensions. Then, the output features of the CNN model are considered as nodes of the graph. Considering the spatial relationship among the patches in the original image, the node from the central patch is connected to the nodes from its neighboring patches to constitute a graph structure, which can be input into a GAT model to learn the relationship among the nodes and update the features. Finally, the output features of GAT can judge whether the central patch contains cracks. Forty original large-sized images are cropped into abundant patches for the training of the CNN-GAT model. With the use of a sliding window technique, the trained CNN-GAT model is capable of finding the patches containing cracks in the test images with large sizes. From the test results, the location and the size of the cracks are exhibited, which indicates that the proposed approach is effective for crack identification in bridge structures.
format Article
id doaj-art-0f54745b79a8447e87c22f2adfa36947
institution OA Journals
issn 2076-3417
language English
publishDate 2025-05-01
publisher MDPI AG
record_format Article
series Applied Sciences
spelling doaj-art-0f54745b79a8447e87c22f2adfa369472025-08-20T02:33:43ZengMDPI AGApplied Sciences2076-34172025-05-011510545210.3390/app15105452Crack Identification for Bridge Condition Monitoring Combining Graph Attention Networks and Convolutional Neural NetworksFeiyu Chen0Tong Tong1Jiadong Hua2Chun Cui3713 Research Institute of CSSC, Zhengzhou 450015, ChinaSchool of Reliability and Systems Engineering, Beihang University, Xueyuan Road No. 37, Haidian District, Beijing 100191, ChinaSchool of Reliability and Systems Engineering, Beihang University, Xueyuan Road No. 37, Haidian District, Beijing 100191, ChinaSchool of Reliability and Systems Engineering, Beihang University, Xueyuan Road No. 37, Haidian District, Beijing 100191, ChinaOrthotropic steel box girders and steel bridge decks are commonly applied to bridges. Because of the coupling of original defects and alternating forces, fatigue cracks are likely to appear in the structures. In order to ensure the life span of bridges, methods for automatic crack identification are needed. In this paper, we present a novel approach for crack detection and bridge condition monitoring by integrating convolutional neural networks (CNNs) with graph attention networks (GATs). At first, the original large-sized images are divided into small-sized patches, and these patches are input into a CNN architecture to extract features by decreasing dimensions. Then, the output features of the CNN model are considered as nodes of the graph. Considering the spatial relationship among the patches in the original image, the node from the central patch is connected to the nodes from its neighboring patches to constitute a graph structure, which can be input into a GAT model to learn the relationship among the nodes and update the features. Finally, the output features of GAT can judge whether the central patch contains cracks. Forty original large-sized images are cropped into abundant patches for the training of the CNN-GAT model. With the use of a sliding window technique, the trained CNN-GAT model is capable of finding the patches containing cracks in the test images with large sizes. From the test results, the location and the size of the cracks are exhibited, which indicates that the proposed approach is effective for crack identification in bridge structures.https://www.mdpi.com/2076-3417/15/10/5452crack identificationconvolutional neural networksgraph attention networksmachine visionbridge condition monitoring
spellingShingle Feiyu Chen
Tong Tong
Jiadong Hua
Chun Cui
Crack Identification for Bridge Condition Monitoring Combining Graph Attention Networks and Convolutional Neural Networks
Applied Sciences
crack identification
convolutional neural networks
graph attention networks
machine vision
bridge condition monitoring
title Crack Identification for Bridge Condition Monitoring Combining Graph Attention Networks and Convolutional Neural Networks
title_full Crack Identification for Bridge Condition Monitoring Combining Graph Attention Networks and Convolutional Neural Networks
title_fullStr Crack Identification for Bridge Condition Monitoring Combining Graph Attention Networks and Convolutional Neural Networks
title_full_unstemmed Crack Identification for Bridge Condition Monitoring Combining Graph Attention Networks and Convolutional Neural Networks
title_short Crack Identification for Bridge Condition Monitoring Combining Graph Attention Networks and Convolutional Neural Networks
title_sort crack identification for bridge condition monitoring combining graph attention networks and convolutional neural networks
topic crack identification
convolutional neural networks
graph attention networks
machine vision
bridge condition monitoring
url https://www.mdpi.com/2076-3417/15/10/5452
work_keys_str_mv AT feiyuchen crackidentificationforbridgeconditionmonitoringcombininggraphattentionnetworksandconvolutionalneuralnetworks
AT tongtong crackidentificationforbridgeconditionmonitoringcombininggraphattentionnetworksandconvolutionalneuralnetworks
AT jiadonghua crackidentificationforbridgeconditionmonitoringcombininggraphattentionnetworksandconvolutionalneuralnetworks
AT chuncui crackidentificationforbridgeconditionmonitoringcombininggraphattentionnetworksandconvolutionalneuralnetworks