Deep Learning for the Detection and Recognition of Rail Defects in Ultrasound B-Scan Images

Abstract

Rail defect detection is crucial to rail operations safety. Addressing the problem of high false alarm rates and missed detection rates in rail defect detection, this paper proposes a deep learning method using B-scan image recognition of rail defects with an improved YOLO (you only look once) V3 algorithm. Specifically, the developed model can automatically position a box in B-scan images and recognize EFBWs (electric flash butt welds), normal bolt holes, BHBs (bolt hole breaks), and SSCs (shells, spalling, or corrugation). First, the network structure of the YOLO V3 model is modified to enlarge the receptive field of the model, thus improving the detection accuracy of the model for small-scale objects. Second, B-scan image data are analyzed and standardized. Third, the initial training parameters of the improved YOLO V3 model are adjusted. Finally, the experiments are performed on 453 B-scan images as the test data set. Results show that the B-scan image recognition model based on the improved YOLO V3 algorithm reached high performance in its precision. Additionally, the detection accuracy and efficiency are improved compared with the original model and the final mean average precision can reach 87.41%.

An improved faster RCNN-based weld ultrasonic atlas defect detection method

“…YOLO target detection algorithm adopts the method in research. 32 A deep learning method of B-scan image recognition of rail defects based on the improved YOLO V3 algorithm, which modifies the network structure of the YOLO V3 model, and uses the Darknet-53 feature extraction network to expand the receptive field of the model. The SSD target detection algorithm adopts the method in research, 33 which is a solid wood board surface defect detection method based on SSD.…”

“…YOLO target detection algorithm adopts the method in research. 32 A deep learning method of B-scan image recognition of rail defects based on the improved YOLO V3 algorithm, which modifies the network structure of the YOLO V3 model, and uses the Darknet-53 feature extraction The overall mean average precision of defect recognition of YOLO algorithm is 86.96%, that of SSD is 88.50%, and the improved Faster RCNN algorithm proposed by others has an overall mean average precision of 89.41%. Compared with the three algorithms, the algorithm in this paper has increased by 6.76%, 5.22%, and 4.31% respectively.…”

Severe rail wear detection with rail running band images

“…Over the past decade, DCNNs have achieved remarkable achievements in various tasks, such as encephalogram classification (Acharya et al., 2018; Nogay & Adeli, 2020), predicting housing prices (Rafiei & Adeli, 2016), and civil engineering structure monitoring (Adeli, 2020; Oh et al., 2017; Rafiei & Adeli, 2018; Rafiei et al., 2017). It has also been extensively used in the railway and other industries to detect railway component defects (Chen et al., 2021; Gao et al., 2022; Guo, Qian, et al., 2021). However, to our best knowledge, a DCNN has not been applied to extract depth information on rail wear.…”

Rail track condition monitoring: a review on deep learning approaches

“…CNN, transfer learning [64] CNN, transfer learning, Bayesian optimization to tune hyperparameters [100] CNN-LSTM [87,103,107] Faster R-CNN [78,88] Faster R-CNN + CNN [73] FastNet, convolutional network-based [120] Fine-grained bilinear CNNs model [70] FCN [119] GAN for CNN [115] Inception-ResNet-v2 & CNN [113] LSTM-RNN [63,71,99] Mask R-CNN…”