The inspection of railway infrastructure faces significant challenges due to heterogeneous environmental conditions and non-uniform illumination patterns, leading to suboptimal detection performance in conventional robotic systems. This study develops a multi-stage image enhancement pipeline incorporating adaptive target segmentation and stereoscopic correspondence matching. A cross-sensor calibration protocol establishes precise spatial coordinates for defect localization through binocular disparity analysis. The proposed framework integrates an enhanced YOLOv5 architecture with context-aware attention modules, developing a hierarchical feature learning architecture that combines pyramidal representation with bidirectional multi-scale feature fusion layers. Experimental validation demonstrates 91.5% precision in fastener absence detection with optimized computational efficiency, indicating substantial improvements in automated rail defect diagnostics compared to baseline systems.
JingGQinXWangH, et al.Developments, challenges, and perspectives of railway inspection robots. Autom ConStruct2022; 138: 1–24.
2.
RyuWLeeS. Development of an inspection robot operating on a single square rail track. J Korea Robot Soc2022; 17: 216–220.
3.
RahmanMLiuHMasriM, et al.A railway track reconstruction method using robotic vision on a mobile manipulator: a proposed strategy. Comput Ind2023; 148: 1–11.
4.
YangHWangYHuJ, et al.Deep learning and machine vision-based inspection of rail surface defects. IEEE Trans Instrum Meas2021; 71: 1–14.
5.
WuYChenPQinY, et al.Automatic railroad track components inspection using hybrid deep learning framework. IEEE Trans Instrum Meas2023; 72: 1–15.
6.
HongboWWeimingQJingW, et al.Detection system of multi-track intelligent inspection robot for urban rail vehicles. Electr Drive Locomot2024; 1: 45–52.
7.
DingYZhuHChenR, et al.An efficient AdaBoost algorithm with the multiple thresholds classification. Appl Sci2022; 12: 23–36.
8.
FanXJiaoXShuaiM, et al.Application research of image recognition technology based on improved SVM in abnormal monitoring of rail fasteners. J Comput Methods Sci Eng2023; 23: 1307–1319.
9.
BiswasRKhanRAIslamS, et al.A novel approach to detect and classify the defective of missing rail anchors in real-time. International Journal of Emerging Technology and Advanced Engineering2016; 6: 270–276.
10.
ManikandanRBalasubramanianMPalanivelS. Machine vision based missing fastener detection in rail track images using SVM classifier. Int J Smart Sens Intell Syst2017; 10: 574–589.
11.
HanCWangSMadanA, et al.Intelligent detection of loose fasteners in railway tracks using distributed acoustic sensing and machine learning. Eng Appl Artif Intell2024; 134: 20–39.
12.
RaguramRChumOPollefeysM, et al.USAC: a universal framework for random sample consensus. IEEE Trans Pattern Anal Mach Intell2012; 35: 2022–2038.
13.
LiLSunRZhaoS, et al.Semantic‐segmentation‐based rail fastener state recognition algorithm. Math Probl Eng2021; 20: 1–15.
14.
ChenJLiuZWangH, et al.Automatic defect detection of fasteners on the catenary support device using deep convolutional neural network. IEEE Trans Instrum Meas2017; 67: 257–269.
15.
GibertXPatelVMChellappaR. Deep multitask learning for railway track inspection. IEEE Trans Intell Transport Syst2016; 18: 153–164.
16.
WeiXYangZLiuY, et al.Railway track fastener defect detection based on image processing and deep learning techniques: a comparative study. Eng Appl Artif Intell2019; 80: 66–81.
17.
AydinISeviMSalurMU, et al.Defect classification of railway fasteners using image preprocessing and alightweight convolutional neural network. Turk J Electr Eng Comput Sci2022; 30: 891–907.
18.
ZhangNDingSZhangJ, et al.An overview on restricted Boltzmann machines. Neurocomputing2018; 275: 1186–1199.
19.
GuoFLiuJQianY, et al.Rail surface defect detection using a transformer-based network. J Indus Inform Intg2024; 38: 117–126.
20.
BrinthaKJoseph JawharS. FOD-YOLO NET: fasteners fault and object detection in railway tracks using deep yolo network. J Intell Fuzzy Syst2024; 46: 8123–8137.
21.
CaiYHeMTaoQ, et al.Fast rail fastener screw detection for vision-based fastener screw maintenance robot using deep learning. Appl Sci2024; 14: 16–30.
22.
TangCZhouGGaoZ, et al.A novel rail inspection robot and fault detection method for the coal mine hoisting system. IEEE Intell Transport Syst Mag2019; 11: 110–121.
23.
IyerSVelmuruganTGandomiAH, et al.Structural health monitoring of railway tracks using IoT-based multi-robot system. Neural Comput Appl2021; 33: 5897–5915.
24.
LvMNiuC. Railway train inspection robot based on intelligent recognition technology. Int J Syst Assur Eng Manag2023; 14: 648–656.
25.
KuoC-CJ. Understanding convolutional neural networks with a mathematical model. J Vis Commun Image Represent2016; 41: 406–413.
26.
NamatēvsI. Deep convolutional neural networks: structure, feature extraction and training. Inf Technol Manag Sci2017; 20: 40–47.
27.
CongSZhouY. A review of convolutional neural network architectures and their optimizations. Artif Intell Rev2023; 56: 1905–1969.
28.
HuiJYangYHuiY, et al.Research on identify matching of object and location algorithm based on binocular vision. J Comput Theor Nanosci2016; 13: 2006–2013.
29.
DemiM. On the gray-level central and absolute central moments and the mass center of the gray-level variability in low-level image processing. Comput Vis Image Understand2005; 97: 180–208.
30.
KangSIwanaBKUchidaS. Complex image processing with less data—document image binarization by integrating multiple pre-trained U-Net modules. Pattern Recogn2021; 109: 25–38.
31.
XiaMMaoQWangG, et al.A subway tunnel image stitching method based on point cloud mapping relationships and high-resolution image. Eng Res Express2024; 6: 025102.
32.
LiJQiuXWeiY, et al.Online rail fastener detection based on YOLO network. Comput Mater Continua (CMC)2022; 12: 59–72.
33.
ZhouXZhangL. SA-FPN: an effective feature pyramid network for crowded human detection. Appl Intell2022; 52: 12556–12568.
34.
MaY. Target tracking and detection based on YOLOv5 algorithm. Appl Comput Eng2023; 16: 75–85.
35.
TervenJCórdova-EsparzaD-MRomero-GonzálezJ-A. A comprehensive review of yolo architectures in computer vision: from yolov1 to yolov8 and yolo-nas. Mach Learn Knowl Extr (2019)2023; 5: 1680–1716.
36.
MukhotiJKulhariaVSanyalA, et al.Calibrating deep neural networks using focal loss. Adv Neural Inf Process Syst2020; 33: 15288–15299.
37.
ZhangZSabuncuM. Generalized cross entropy loss for training deep neural networks with noisy labels. Adv Neural Inf Process Syst2018; 31: 12563–12571.
38.
CuiHLiJHuQ, et al.Real-time inspection system for ballast railway fasteners based on point cloud deep learning. IEEE Access2019; 8: 61604–61614.
