Damage identification method of prestressed concrete beam bridge based on convolutional neural network

Yang, Sanqiang; Yong, Huang

doi:10.1007/s00521-020-05052-w

Cited by 26 publications

(12 citation statements)

References 26 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Other structural modal information, such as natural frequencies and mode shapes (Pathirage et al 2019;Wang et al 2021), are also sensitive to damage. Yang and Huang (2021) introduced the flexibility curvature index that did not need the information of intact structures as the input of a CNN to realize damage identification. Nguyen et al (2020) trained a CNN using the images from the damage index of the gapped smoothing method to classify the damage location in a numerical beam.…”

Section: Damage Scenario Classificationmentioning

confidence: 99%

The application of deep learning in bridge health monitoring: a literature review

Zhang

Wang

et al. 2022

ABEN

View full text Add to dashboard Cite

Along with the advancement in sensing and communication technologies, the explosion in the measurement data collected by structural health monitoring (SHM) systems installed in bridges brings both opportunities and challenges to the engineering community for the SHM of bridges. Deep learning (DL), based on deep neural networks and equipped with high-end computer resources, provides a promising way of using big measurement data to address the problem and has made remarkable successes in recent years. This paper focuses on the review of the recent application of DL in SHM, particularly damage detection, and provides readers with an overall understanding of the missions faced by the SHM of the bridges. The general studies of DL in vibration-based SHM and vision-based SHM are respectively reviewed first. The applications of DL to some real bridges are then commented. A summary of limitations and prospects in the DL application for bridge health monitoring is finally given.

show abstract

Section: Damage Scenario Classificationmentioning

confidence: 99%

The application of deep learning in bridge health monitoring: a literature review

Zhang

Wang

et al. 2022

ABEN

View full text Add to dashboard Cite

show abstract

“…It is notable that except for one, all studies have utilized IMA-GENET as a source dataset. Additionally, pre-trained CNN models are also being utilized for vibration-based damage localization [1,4,69], condition assessment [30], and fault diagnosis [66]. Specific (visual) detection tasks in the damage recognition setting are strongly influenced by various operating conditions, such as surface reflectance, roughness, concrete materials, coatings, and weather phenomena for different components of a bridge [42].…”

Section: Damage Detection Using Deep (Transfer) Learningmentioning

confidence: 99%

Damage detection using in-domain and cross-domain transfer learning

Bukhsh

Jansen

Saeed

2021

Neural Comput & Applic

View full text Add to dashboard Cite

We investigate the capabilities of transfer learning in the area of structural health monitoring. In particular, we are interested in damage detection for concrete structures. Typical image datasets for such problems are relatively small, calling for the transfer of learned representation from a related large-scale dataset. Past efforts of damage detection using images have mainly considered cross-domain transfer learning approaches using pre-trained ImageNet models that are subsequently fine-tuned for the target task. However, there are rising concerns about the generalizability of ImageNet representations for specific target domains, such as for visual inspection and medical imaging. We, therefore, evaluate a combination of in-domain and cross-domain transfer learning strategies for damage detection in bridges. We perform comprehensive comparisons to study the impact of cross-domain and in-domain transfer, with various initialization strategies, using six publicly available visual inspection datasets. The pre-trained models are also evaluated for their ability to cope with the extremely low-data regime. We show that the combination of cross-domain and in-domain transfer persistently shows superior performance specially with tiny datasets. Likewise, we also provide visual explanations of predictive models to enable algorithmic transparency and provide insights to experts about the intrinsic decision logic of typically black-box deep models.

show abstract

“…Therefore, several studies adopt standard architectures-such as VGG16, Inception-v3, or ResNet50 pre-trained models with IMAGENET representations-for the detection of cracks [49,44,56,26,42], potholes [4,28], spalls [50], and multiple other damages including corrosion, seapage, and exposed bars [32,20,58,59,55,12]. Additionally, pre-trained CNN models are also being utilized for vibration-based damage localization [1,3,51], condition assessment [22] and fault diagnosis [48].…”

Section: Damage Detection Using Deep (Transfer) Learningmentioning

confidence: 99%

Damage detection using in-domain and cross-domain transfer learning

Bukhsh,

Jansen,

Saeed

2021

Preprint

View full text Add to dashboard Cite

We investigate the capabilities of transfer learning in the area of structural health monitoring. In particular, we are interested in damage detection for concrete structures. Typical image datasets for such problems are relatively small, calling for the transfer of learned representation from a related large-scale dataset. Past efforts of damage detection using images have mainly considered cross-domain transfer learning approaches using pre-trained IMAGENET models that are subsequently fine-tuned for the target task. However, there are rising concerns about the generalizability of IMAGENET representations for specific target domains, such as for visual inspection and medical imaging. We, therefore, propose a combination of in-domain and cross-domain transfer learning strategies for damage detection in bridges. We perform comprehensive comparisons to study the impact of crossdomain and in-domain transfer, with various initialization strategies, using six publicly available visual inspection datasets. The pre-trained models are also evaluated for their ability to cope with the extremely low-data regime. We show that the combination of cross-domain and in-domain transfer persistently shows superior performance even with tiny datasets. Likewise, we also provide visual explanations of predictive models to enable algorithmic transparency and provide insights to experts about the intrinsic decision-logic of typically black-box deep models.Preprint. Under review.

show abstract

Damage identification method of prestressed concrete beam bridge based on convolutional neural network

Cited by 26 publications

References 26 publications

The application of deep learning in bridge health monitoring: a literature review

The application of deep learning in bridge health monitoring: a literature review

Damage detection using in-domain and cross-domain transfer learning

Damage detection using in-domain and cross-domain transfer learning

Contact Info

Product

Resources

About