A robust bridge rivet identification method using deep learning and computer vision

Jiang, Tengjiao; Frøseth, Gunnstein T.; Rönnquist, Anders

doi:10.1016/j.engstruct.2023.115809

Cited by 17 publications

(5 citation statements)

References 29 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…There are many research studies that utilize CNN models for edge detection. For instance, CNNs have been extensively employed in bridge rivet identification tasks, where they demonstrate remarkable performance [ 6 ] Both of the above-mentioned model types typically employ large backbone networks as pre-training models, which often results in high computational costs. To address this issue, some research works have made improvements.…”

Section: Related Workmentioning

confidence: 99%

“…Edge detection, as a traditional computer vision task, aims to identify prominent changes in brightness or discontinuous regions in an image, making it an important research area in feature extraction. It provides fundamental information for many advanced visual tasks, such as image segmentation [ 1 , 2 ], contour extraction [ 3 ], object detection [ 4 , 5 , 6 ], and 2D object recognition [ 7 ]. With the advancement of deep learning, new domains, such as medical image analysis and remote sensing, have emerged, which often require an edge detection system.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Edge Detection via Fusion Difference Convolution

Yin

Wang

Fan

et al. 2023

Sensors

View full text Add to dashboard Cite

Edge detection is a crucial step in many computer vision tasks, and in recent years, models based on deep convolutional neural networks (CNNs) have achieved human-level performance in edge detection. However, we have observed that CNN-based methods rely on pre-trained backbone networks and generate edge images with unwanted background details. We propose four new fusion difference convolution (FDC) structures that integrate traditional gradient operators into modern CNNs. At the same time, we have also added a channel spatial attention module (CSAM) and an up-sampling module (US). These structures allow the model to better recognize the semantic and edge information in the images. Our model is trained from scratch on the BIPED dataset without any pre-trained weights and achieves promising results. Moreover, it generalizes well to other datasets without fine-tuning.

show abstract

Section: Related Workmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Edge Detection via Fusion Difference Convolution

Yin

Wang

Fan

et al. 2023

Sensors

View full text Add to dashboard Cite

show abstract

“…In the past few decades, with the continuous emergence and rapid development of emerging technologies, artifcial intelligence (AI) and deep learning (DL) techniques have been gradually applied to bridge construction, operation, and maintenance [15][16][17][18][19]. For instance, Bae et al [20] developed an automated crack detection method by combining deep super-resolution crack networks and deep learning models.…”

Section: Introductionmentioning

confidence: 99%

Pixel‐Level Crack Identification for Bridge Concrete Structures Using Unmanned Aerial Vehicle Photography and Deep Learning

Song,

Liu,

Yuan

2024

Structural Control and Health Monitoring

View full text Add to dashboard Cite

Traditional manual inspection technology has the problems of high risk, low efficiency, and being time-consuming in bridge safety management. The unmanned aerial vehicle (UAV)-based detection technology is widely used in bridge structure safety monitoring. However, the existing deep learning-based concrete crack identification method has great limitations in dealing with complex background and tiny cracks in bridge structures. To address these problems, this study designs a crack pixel-level high-performance segmentation model for bridge concrete cracks that is suitable for UAV detection scenarios using machine vision (MV) and deep learning (DL) algorithms. First, considering the high requirements for the computing performance of the MV-based model for UAV-based detection, the ResNet-18-based lightweight convolutional neural network is used to represent the traditional large-scale backbone network of the pyramid scene parsing network (PSPNet) to develop a high-performance crack automatic identification model. Then, considering that bridge concrete cracks have the characteristics of subtle shapes and complex backgrounds, the spatial position self-attention module is inserted into the PSPNet to improve its detection accuracy. A concrete bridge is used for the case study, and a dataset of cracks in bridge concrete structures collected by UAVs is constructed and used for model training. The experimental results show that the loss function of the developed method in the training process results in a smooth decline, and the developed algorithm achieves the evaluation indicators of 0.9008 precision, 0.8750 recall, 0.8820 accuracy, and 0.9012 IOU on the bridge concrete crack dataset, which are significantly higher than other state-of-the-art baseline methods. In addition, four common UAV bridge detection scenarios, including low light, complex crack forms, high background roughness, and complex background scenes, are used to further test the crack detection ability of the developed crack identification model. The experimental results show that the proposed crack identification method can effectively overcome interference and real-size pixel-level segmentation of crack morphology. In addition, it also achieved a detection efficiency of 35.04 FPS, which shows that the real-time detection ability of the method has good applicability in the UAV detection scene.

show abstract

“…In recent decades, a noteworthy shift in research focus has been observed towards Structural Health Monitoring (SHM), with a primary emphasis on predicting structural conditions to inform decision-making through data utilization (Artagan et al, 2020;Jiang et al, 2023). The application of advanced sensor technologies has proven beneficial for the continuous and reliable inspection of critical railway infrastructure assets, including tracks.…”

Section: Introductionmentioning

confidence: 99%

Identification of optimal accelerometer placement on trains for railway switch wear monitoring via multibody simulation

Hu,

Lau,

Dai

et al. 2024

Front. Built Environ.

View full text Add to dashboard Cite

Accelerometers play a crucial role in the railway industry, especially in track monitoring. Traditionally, they are placed on the railway tracks or often on bridges to monitor the health and condition of the infrastructure. Recently, there has been an increased focus on using regular trains to monitor the condition of railway infrastructure. Often, the sensors are placed based on certain assumptions without much scientific evidence or support. This paper utilizes the multibody simulation software GENSYS to identify the optimal placement of accelerometers on a passenger train for monitoring railway switch wear. Switch wear profiles were generated systematically and used as input for the simulations, studying acceleration at a total of 93 locations distributed among the wheelsets, bogies, and carbody. Based on both time and frequency domain analyses, optimal sensor locations were identified, generally close to the first bogie or wheelset at the leading carbody. Accelerations generated by the wheelset passing the switch can also be captured in the carbody, but it is important to note that these are several orders lower in magnitude compared to the acceleration on the wheelset. If accelerometers are to be placed in the carbody, correct sensitivity must be chosen, and a high-pass filter should be applied to capture the acceleration signals associated with switch wear. The study confirms that there is a direct correlation between the depth of switch wear and the magnitude of the acceleration. It remains effective even under various curve radii and train speeds.

show abstract

A robust bridge rivet identification method using deep learning and computer vision

Cited by 17 publications

References 29 publications

Edge Detection via Fusion Difference Convolution

Edge Detection via Fusion Difference Convolution

Pixel‐Level Crack Identification for Bridge Concrete Structures Using Unmanned Aerial Vehicle Photography and Deep Learning

Identification of optimal accelerometer placement on trains for railway switch wear monitoring via multibody simulation

Contact Info

Product

Resources

About