The character recognition of angle steel in factory environments is often challenged by factors such as noise, rust, and varying lighting conditions. To address this, we explore the application of the YOLOX model and propose an enhanced version, T-YOLOX. By integrating the Ghost module and FReLU activation, T-YOLOX achieves model lightweighting and improves the ability to capture character features, thereby enhancing recognition speed and accuracy. The effectiveness of these improvements was validated through ablation experiments. Moreover, we conducted comparative experiments between T-YOLOX, several YOLO models, and other mainstream object detection algorithms, demonstrating that T-YOLOX performs better across multiple key evaluation metrics. Furthermore, we developed two practical applications: an embedded recognition system for production lines and a mobile handheld system for real-time recognition in outdoor and distributed scenarios, both supporting seamless MES integration. These results extend the application of YOLOX in industrial settings and provide robust technical support for angle steel character recognition in steel tower manufacturing.
ShuXHZhangSGLiY, et al. An anomaly detection method based on random convolutional kernel and isolation forest for equipment state monitoring. Eksploat Niezawodn2022; 24(4): 758–770.
2.
MelnikRKoziakSDižoJ, et al. Feasibility study of a rail vehicle damper fault detection by artificial neural networks. Eksploat Niezawodn2023; 25(1): 1–11.
3.
JiaMXLüYPXuLF, et al. The influence of friction on the texture formation of a IF steel during hot rolling in the ferrite region. Steel Res Int2013; 84(8): 761–765.
4.
PanHYZhangXHWuJ, et al. Sustainability evaluation of a steel production system in China based on emergy. J Clean Prod2016; 112: 1498–1509.
5.
LuoYChenSWHeXJ, et al. Alphanumeric character recognition based on BP neural network classification and combined features. Int J Comput Intell Syst2013; 6(6): 1108–1115.
6.
BoulidYSouharAElkettaniMY.Detection of text lines of handwritten Arabic manuscripts using Markov decision processes. Int J Interact Multimed Artif Intell2016; 4(1): 31–36.
7.
HinamiRLiuXHChibaN, et al. Bidirectional extraction and recognition of scene text with layout consistency. Int J Doc Anal Recognit2016; 19(2): 83–98.
8.
ZhaoYBCongSBuXY.Study on automated guided vehicle collision avoidance mechanism with external computer vision. Proc Inst Mech Eng Part B: J Eng Manuf. Epub ahead of print 15January2025. DOI: 10.1177/09544054241310336.
9.
DuanJGZhuangLWZhangQL, et al. Vision-based robotic grasping using faster R-CNN-GRCNN dual-layer detection mechanism. Proc Inst Mech Eng Part B: J Eng Manuf2025; 239(6–7): 950–964.
YangJSunLWangG.Improving the underwater object detection algorithm for YOLOv5. In: Proceedings of the 3rd international conference on electronic information engineering and data processing, Kuala Lumpur, Malaysia, 2024, vol. 13184. SPIE.
12.
YingSSLiMHFangJH, et al. Recognition method study on stamping strings of angle steel parts for intelligent manufacturing of power transmission towers. J Eng Des2025; 1–28. DOI:10.1080/09544828.2025.2506293
13.
YingSSFangJHTangSZ, et al. Improved YOLOV5 angle embossed character recognition by multiscale residual attention with selectable clustering. Electronics2024; 13(13): 2435.
14.
AshrafIHurSKimG, et al. Analyzing performance of YOLOx for detecting vehicles in bad weather conditions. Sensors2024; 24(2): 522.
15.
TsaiCYWangRYChiuYC.SW-YOLOX: a YOLOX-based real-time pedestrian detector with shift window-mixed attention mechanism. Neurocomputing2024; 606: 128357.
16.
ZhouQYuCH.Object detection made simpler by eliminating heuristic NMS. IEEE Trans Multimedia2023; 25: 9254–9262.
17.
ChenDBLüLYShangMS, et al. Identifying influential nodes in complex networks. Physica A Stat Mech Appl2012; 391(4): 1777–1787.
18.
ArunDRColumbusCCBhuvaneshN, et al. Smart crowd monitoring system using IoT-based YOLO-GHOST. Rev Roum Sci Tech Ser Electro2024; 69(3): 345–350.
19.
KothalaLPJonnalaPGunturSR.Localization of mixed intracranial hemorrhages by using a ghost convolution-based YOLO network. Biomed Signal Process Control2023; 80: 104378.
20.
KawakamiMHirataKFuruyaS, et al. Development of combination methods for detecting malignant uptakes based on physiological uptake detection using object detection with PET-CT MIP images. Front Med2020; 7: 616746.
21.
MathiasADhanalakshmiSKumarR.Occlusion aware underwater object tracking using hybrid adaptive deep SORT-YOLOv3 approach. Multimed Tools Appl2022; 81(30): 44109–44121.
22.
NguyenDDNguyenDTLeMT, et al. FPGA-SoC implementation of YOLOv4 for flying-object detection. J Real-Time Image Process2024; 21(3): 63.
23.
AtreyJRegunathanRRajasekaranR.Real-world application of face mask detection system using YOLOv6. Int J Crit Infrastruct2024; 20(3): 216–240.
24.
YadavPGuptaNSharmaPK.Robust weapon detection in dark environments using Yolov7-DarkVision. Digit Signal Prog2024; 145: 104342.
25.
ZhaoZPHeCZhaoGM, et al. RA-YOLOX: re-parameterization align decoupled head and novel label assignment scheme based on YOLOX. Pattern Recognit2023; 140: 109579.
26.
MoonJJeonMJeongS, et al. RoMP-transformer: rotational bounding box with multi-level feature pyramid transformer for object detection. Pattern Recognit2024; 147: 110067.
27.
TanMXPangRMLeQ.EfficientDet: scalable and efficient object detection. In: IEEE/CVF conference on computer vision and pattern recognition (CVPR), Seattle, WA, 13–19 June 2020, pp.10778–10787. IEEE.
28.
LinTYGoyalPGirshickR, et al. Focal loss for dense object detection. IEEE Trans Pattern Anal Mach Intell2020; 42(2): 318–327.
29.
SaidYAtriMAlbaharMA, et al. Indoor signs detection for visually impaired people: navigation assistance based on a lightweight anchor-free object detector. Int J Environ Res Public Health2023; 20(6): 5011.
30.
LawHDengJ.CornerNet: detecting objects as paired keypoints. Int J Comput Vis2020; 128(3): 642–656.
31.
TianZShenCHChenH, et al. FCOS: a simple and strong anchor-free object detector. IEEE Trans Pattern Anal Mach Intell2022; 44(4): 1922–1933.
32.
MattinaMBenzinouANasreddineK, et al. An efficient anchor-free method for pig detection. IET Image Process2023; 17(2): 613–626.
33.
BenjinZJianfengWZhengkaiJ, et al. AutoAssign: differentiable label assignment for dense object detection. arXiv. 2020. DOI: 10.48550/arXiv.2007.03496.
34.
GeZLiuSTLiuZM, et al. OTA: optimal transport assignment for object detection. In: IEEE/CVF conference on computer vision and pattern recognition (CVPR), Nashville, TN, 19–25 June 2021, pp. 303–312. IEEE.
