A lightweight foreign object detection framework for high-speed trains with adaptive illumination enhancement

Abstract

The detection of foreign objects in key components of high-speed trains is critical for railway safety, but existing methods struggle under low-light conditions, complex backgrounds, and small objects. To address these issues, we propose an efficient detection framework integrating an adaptive brightness enhancement network (ABEN) and a lightweight train foreign object detection network (LTFD-Net). ABEN adaptively enhances images according to their illumination, improving clarity across multiple objects and backgrounds while ensuring real-time processing. LTFD-Net combines a lightweight backbone with a multi-dimensional feature enhancement module, capturing multi-scale and contextual features to accurately detect small and complex defects with minimal computational overhead. To support realistic evaluation, we introduce the high-speed train foreign object detection (HTFD) dataset with 3,904 annotated images across five key components. Experiments show that the integrated framework achieves 85.4% mean average precision (mAP) and 112 frames per second (FPS) on HTFD, surpassing state-of-the-art methods. Independently, LTFD-Net reaches 78.3% mAP and 97 FPS on NEU-DET, demonstrating preliminary generalization capability without illumination enhancement. This work provides an efficient and practical solution for real-time foreign object detection under challenging railway conditions.

Keywords

Foreign object detection / Low-light conditions / Adaptive brightness enhancement / Lightweight detection network

References

[1] Wei X, Jiang S, Li Y, Li C, Jia L, Li Y. Defect detection of pantograph slide based on deep learning and image processing technology. IEEE Trans Intell Transp Syst, 2020, 21: 947-958

[2] Wen Z, Shao T, Li P, Kai M. Defects detection and segmentation of operation safety image of EMU based on convolutional neural network. J China Railway Soc, 2019, 41: 76-83

[3] Zhao Y, Lv W, Xu S et al (2024) DETRs beat YOLOs on real-time object detection. In: 2024 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). June 16–22, Seattle, pp 16965–16974

[4] Liu S, Qi L, Qin H et al (2018) Path aggregation network for instance segmentation. In: 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition. June 18–23, Salt Lake City, pp 8759–8768

[5] Zhang Y, Qin N, Huang D et al (2022) Precise diagnosis of unknown fault of high-speed train bogie using novel FBM-net. IEEE Trans Instrum Meas 71:3526811

[6] Niu M, Song K, Huang L, Wang Q, Yan Y, Meng Q. Unsupervised saliency detection of rail surface defects using stereoscopic images. IEEE Trans Ind Inf, 2021, 17: 2271-2281

[7] Guo F, Liu J, Qian Y, Xie Q. Rail surface defect detection using a transformer-based network. J Ind Inf Integr, 2024, 38 ArticleID: 100584

[8] Hong Z, Xiao H, Li S, Qin Y, Yu C, Zhang L, Zhao K. FDSP-HRID: few-shot detector with self-supervised pretraining for high-speed rail infrastructure defects. IEEE Trans Instrum Meas, 2024, 73: 12

[9] Li B, Zhang L, Wang H, Liu Y, Chen J, Zhao Q. Toward reliable high-speed railway pantograph-catenary system state detection: Multitask deep neural networks with runtime reliability monitoring. IEEE Trans Instrum Meas, 2024, 73: 11

[10] Hong Z, Xiao H, Li S et al (2023) Few-shot object detection for high-speed rail infrastructure defects. In: China Automation Congress (CAC). November 17–19, Chongqing, pp 9344–9349

[11] Zhang Q, Nie Y, Zheng WS (2019) Dual illumination estimation for robust exposure correction. arXiv preprint arXiv:1910.13688

[12] Guo X, Li Y, Ling H. LIME: low-light image enhancement via illumination map estimation. IEEE Trans Image Process, 2017, 26: 982-993

[13] Jeon JJ, Park JY, Eom IK. Low-light image enhancement using gamma correction prior in mixed color spaces. Pattern Recognit, 2024, 146 ArticleID: 110001

[14] Wu W, Weng J, Zhang P et al (2022) URetinex-net: retinex-based deep unfolding network for low-light image enhancement. In: IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). June 18–24, New Orleans, pp 5891–5900

[15] Cui Z, Li K, Lu G et al (2022) You only need 90K parameters to adapt light: a lightweight transformer for image enhancement and exposure correction. In: Proceedings of the AAAI Conference on Artificial Intelligence (AAAI). pp 792–800

[16] Goodfellow IJ, Pouget-Abadie J, Mirza M et al (2014) Generative adversarial nets. In: Advances in Neural Information Processing Systems (NeurIPS), pp 2672–2680

[17] Solunke BR, Gengaje SR (2023) A review on traditional and deep learning based object detection methods. In: International Conference on Emerging Smart Computing and Informatics (ESCI). March 1–3, Pune, pp 1–7

[18] Bochkovskiy A, Wang CY, Liao HYM (2020) YOLOv4: Optimal speed and accuracy of object detection. arXiv preprint arXiv:2004.10934

[19] Varghese R, Sambath M (2024) YOLOv8: a novel object detection algorithm with enhanced performance and robustness. In: International Conference on Advances in Data Engineering and Intelligent Computing Systems (ADICS). April 18–19, Chennai, pp 1–6

[20] Wang A, Chen H, Liu L et al (2024) YOLOv10: Real-time end-to-end object detection. arXivpreprint arXiv:2405.14458

[21] Khanam R, Hussain M (2024) YOLOv11: An overview of the key architectural enhancements. arXiv preprint arXiv:2410.17725

[22] Ren S, He K, Girshick R, Sun J. Faster R-CNN: towards real-time object detection with region proposal networks. IEEE Trans Pattern Anal Mach Intell, 2017, 39: 1137-1149

[23] Carion N, Massa F, Synnaeve G et al (2020) End-to-end object detection with transformers. arXiv preprint arXiv:2005.12872

[24] Chollet F (2017) Xception: deep learning with depthwise separable convolutions. In: IEEE Conference on Computer Vision and Pattern Recognition (CVPR). July 21–26, Honolulu, pp 1800–1807

[25] Yang L, Zhang RY, Li L et al (2021) SimAM: A simple, parameter-free attention module for convolutional neural networks. In: Proceedings of the 38th International Conference on Machine Learning (ICML). July 18–24, pp 11863–11874

[26] Jocher G, Chaurasia A, Qiu J et al (2023) YOLOv5 [software]. GitHub repository. https://github.com/ultralytics/yolov5

[27] Han K, Wang Y, Tian Q et al (2020) GhostNet: more features from cheap operations. In: IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). June 13–19, Seattle, pp 1577–1586

[28] Lin H, Liu B, Li J, Li X. HFD-Net: A benchmark framework of foreign object detection for high-speed train. Pattern Recogn, 2026, 173 ArticleID: 112932

[29] He Y, Song K, Meng Q, Yan Y. An end-to-end steel surface defect detection approach via fusing multiple hierarchical features. IEEE Trans Instrum Meas, 2020, 69: 1493-1504

[30] Lin TY, Goyal P, Girshick R, He K, Dollar P. Focal loss for dense object detection. IEEE Trans Pattern Anal Mach Intell, 2020, 42: 318-327

[31] Liu W, Anguelov D, Erhan D et al (2016) SSD: single shot MultiBox detector. In: Proceedings of the European Conference on Computer Vision. October 11–14, Amsterdam, pp 21–37

[32] Zhu X, Su W, Lu L et al (2020) Deformable DETR: Deformable transformers for end-to-end object detection. arXiv:2010.04159

[33] He Y, Li J. TSRes-YOLO: an accurate and fast cascaded detector for waste collection and transportation supervision. Eng Appl Artif Intell, 2023, 126 ArticleID: 106997

[34] Wang Q, Wu B, Zhu P et al (2020) ECA-net: efficient channel attention for deep convolutional neural networks. In: IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). June 13–19, Seattle, pp 11531–11539

[35] Dong J, Wang N, Fang H, Shen Y, Li B, Di D, Zhai K. CBAM-optimized automatic segmentation and reconstruction system for monocular images with asphalt pavement potholes. IEEE Trans Intell Transp Syst, 2024, 25(8): 10313-10330

[36] Lau KW, Po LM, Rehman YAU. Large separable kernel attention: rethinking the large kernel attention design in CNN. Expert Syst Appl, 2024, 236 ArticleID: 121352

[37] Cheng X, Yu J (2021) RetinaNet with difference channel attention and adaptively spatial feature fusion for steel surface defect detection. IEEE Trans Instrum Meas 70:2503911

[38] Liu G, Chu M, Gong R, Zheng Z. Global attention module and cascade fusion network for steel surface defect detection. Pattern Recogn, 2025, 158 ArticleID: 110979

Funding

National Key Research and Development Program of China(2024YFC3308305)

National Natural Science Foundation of China(U2133218)

Fundamental Research Funds for the Central Universities of China(No.FRF-MP-19-007 and No. FRF-TP-20-065A1Z)

RIGHTS & PERMISSIONS

The Author(s)

Just Accepted

This article has successfully passed peer review and final editorial review, and will soon enter typesetting, proofreading and other publishing processes. The currently displayed version is the accepted final manuscript. The officially published version will be updated with format, DOI and citation information upon launch. We recommend that you pay attention to subsequent journal notifications and preferentially cite the officially published version. Thank you for your support and cooperation.