Pandrol track fastener defect detection based on local convolutional neural networks

Ma, Anqi; Lv, Zhaomin; Chen, Xingjie; Li, Liming; Qiu, Yuzhuo; Zheng, Shibao; Chai, Xiaodong

doi:10.1177/0959651820953679

Cited by 9 publications

(18 citation statements)

References 44 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…The railway fastener system helps to fix rails on sleepers or track slabs and provides elasticity for the rails [74,75]. There are many types of fastener system all over the world, such as 102 and 8 K types from Japan, Nabla from French [76], RST and VOSSLOH 300 from German [77,78], Pandrol series from the UK [79] and WJ series from China [80]. Their components and structures are different, but they can be regarded as comprising three parts: withholding parts (e.g., clips, clamps), jointing parts (e.g., dog spikes, screw spikes and dowels) and elastic pads (rail pads).…”

Section: Fastenersmentioning

confidence: 99%

State-of-the-Art Review on Additive Manufacturing Technology in Railway Infrastructure Systems

Kaewunruen

2021

J. Compos. Sci.

View full text Add to dashboard Cite

Additive manufacturing technologies, well known as three-dimensional printing (3DP) technologies, have been applied in many industrial fields, including aerospace, automobiles, shipbuilding, civil engineering and nuclear power. However, despite the high material utilization and the ability to rapidly construct complex shaped structures of 3D printing technologies, the application of additive manufacturing technologies in railway track infrastructure is still at the exploratory stage. This paper reviews the state-of-the-art research of additive manufacturing technologies related the railway track infrastructure and discusses the challenges and prospects of 3D printing technology in this area. The insights will not only help the development of 3D printing technologies into railway engineering but also enable smarter railway track component design and improve track performance and inspection strategies.

show abstract

Section: Fastenersmentioning

confidence: 99%

State-of-the-Art Review on Additive Manufacturing Technology in Railway Infrastructure Systems

Kaewunruen

2021

J. Compos. Sci.

View full text Add to dashboard Cite

show abstract

“…Flow chart of detection of defects in Pandrol track fasteners based on LDFFN is shown in Figure 3. (1) Segmentation rate (SR): N O is represented as the number of original images, and N S is represented as the number of pictures successfully segmented [41].…”

Section: Track Fastenermentioning

confidence: 99%

“…(2) False alarm rate (FAR), missed alarm rate (MAR), and error rate (ER): D A A indicates that the abnormal track fastener detection result is abnormal; D N A indicates that the abnormal track fastener test result is normal; D N N indicates that the normal track fastener test result is normal; D A N indicates that the normal track fastener test result is abnormal [41].…”

Section: Experiments and Applicationsmentioning

confidence: 99%

Defect Detection of Pandrol Track Fastener Based on Local Depth Feature Fusion Network

Chen

et al. 2021

Complexity

Self Cite

View full text Add to dashboard Cite

There are three main problems in track fastener defect detection based on image: (1) The number of abnormal fastener pictures is scarce, and supervised learning detection model is difficult to establish. (2) The potential data features obtained by different feature extraction methods are different. Some methods focus on edge features, and some methods focus on texture features. Different features have different detection capabilities, and these features are not effectively fused and utilized. (3) The detection of the track fastener clip will be interfered by the track fastener bolt subimage. Aiming at the above three problems, a method for track fastener defects detection based on Local Deep Feature Fusion Network (LDFFN) is proposed. Firstly, the track fastener image segmentation method is used to obtain the track fastener clip subimage, which can effectively reduce the interference of bolt subimage features on the track fastener clip detection. Secondly, the edge features and texture features of track fastener clip subimages are extracted by Autoencoder (AE) and Restricted Boltzmann Machine (RBM), and the features are fused. Finally, the similarity measurement method Mahalanobis Distance (MD) is used to detect defects in track fasteners. The effectiveness of the proposed method is verified by real Pandrol track fastener images.

show abstract

“…Gibert et al [24] used a customized fully convolutional network to extract the highly abstract features of fasteners and identify fastener types and then utilized customized support vector machines to classify the state of fasteners for various types of fasteners. Ma et al [25] cropped out the bolt area subimages that were not related to the identification of the fastener state on the fastener area image and then used the CNN network for classification. rough this approach, the accuracy rate is improved compared to that with the classification directly in the fastener area.…”

Section: Introductionmentioning

confidence: 99%

A Defect Detection Method for Rail Surface and Fasteners Based on Deep Convolutional Neural Network

Zheng

et al. 2021

Computational Intelligence and Neuroscience

Self Cite

View full text Add to dashboard Cite

As a result of long-term pressure from train operations and direct exposure to the natural environment, rails, fasteners, and other components of railway track lines inevitably produce defects, which have a direct impact on the safety of train operations. In this study, a multiobject detection method based on deep convolutional neural network that can achieve nondestructive detection of rail surface and fastener defects is proposed. First, rails and fasteners on the railway track image are localized by the improved YOLOv5 framework. Then, the defect detection model based on Mask R-CNN is utilized to detect the surface defects of the rail and segment the defect area. Finally, the model based on ResNet framework is used to classify the state of the fasteners. To verify the robustness and effectiveness of our proposed method, we conduct experimental tests using the ballast and ballastless railway track images collected from Shijiazhuang-Taiyuan high-speed railway line. Through a variety of evaluation indexes to compare with other methods using deep learning algorithms, experimental results show that our method outperforms others in all stages and enables effective detection of rail surface and fasteners.

show abstract

Pandrol track fastener defect detection based on local convolutional neural networks

Cited by 9 publications

References 44 publications

State-of-the-Art Review on Additive Manufacturing Technology in Railway Infrastructure Systems

State-of-the-Art Review on Additive Manufacturing Technology in Railway Infrastructure Systems

Defect Detection of Pandrol Track Fastener Based on Local Depth Feature Fusion Network

A Defect Detection Method for Rail Surface and Fasteners Based on Deep Convolutional Neural Network

Contact Info

Product

Resources

About