Railway fasteners play a crucial role in connecting the track to the sleepers, and their performance is vital for the stability of the track structure and the safety of train operations. Traditional detection methods primarily rely on human experience, which can be inefficient and susceptible to human error. This paper proposes a novel method for railway fastener detection based on YOLOv8s. By incorporating the DAS dynamic convolution module and EMA attention mechanism modules, the method effectively filters out irrelevant background information, allowing the model to focus on critical areas within the image without increasing computational overhead. Experimental results indicate that the improved model demonstrates superior recognition performance for railway fasteners in complex environments. This research not only significantly improve the efficiency and accuracy of detecting the integrity of railway fasteners, but also ensure the safety and reliability of railway transportation.


    Access

    Check access

    Check availability in my library

    Order at Subito €


    Export, share and cite



    Title :

    Track Fastener Detection Method Based on Improved Yolov8s Algorithm


    Contributors:
    Zhu, Jianpeng (author) / Zhang, Haigang (author) / Wu, Lei (author) / Yang, Minglai (author)


    Publication date :

    2024-11-15


    Size :

    995090 byte




    Type of media :

    Conference paper


    Type of material :

    Electronic Resource


    Language :

    English



    Vehicle detection research based on improved YOLOv8s

    Yu, Qing / Song, Anjun | SPIE | 2025


    Grassland Fire Detection and Localization of UAV Based on Optimized YOLOv8s

    Ma, Junlin / Wang, Fangping / Liu, Bin et al. | IEEE | 2025



    Deep Learning Network Model for Road Pavement Damage Detection Based on YOLOv8s-GES

    Wang, Niannian / Gao, Ziqiao / Fang, Hongyuan et al. | Transportation Research Record | 2025


    Track Fastener Defect Detection Based on Local Convolutional Neural Networks

    Chen, Xingjie / Ma, Anqi / Lv, Zhaomin et al. | TIBKAT | 2021