The embodiment of the invention discloses a rail fastener service state abnormity identification method, device and system based on deep learning. The device comprises the steps that equipment is fixed to the bottom of a rail train through a tool framework; obtaining service state data of the track fastener through a high-speed photography recognition technology; transmitting the service state data to a detection system based on deep learning by using a video line; a control circuit is connected to reach a driver control room to control start and stop and receive a detection result; the method comprises the following steps: acquiring service state data of a track fastener; respectively inputting the operation data into a plurality of fault detection models and a normal detection model, the plurality of fault detection models and the normal detection model being self-encoders based on deep learning, the plurality of fault detection models are respectively used for coding to obtain different fault types and different fault type combination characteristics of the track fastener, and the normal detection model is used for coding to obtain normal characteristics of the track fastener; if the output features of the normal detection model are not matched with the features in a normal feature library, comparing the output features of the plurality of fault detection models with feature libraries of corresponding fault types or combinations; and identifying the fault of the track fastener according to the comparison result. According to the embodiment, automatic identification of the track fastener fault is realized, and the identification precision is improved.

    本发明实施例公开了一种基于深度学习的轨道扣件服役状态异常识别方法、设备和系统。其中,装置包括:利用工装构架将设备固定于轨道列车底部;通过高速摄影识别技术获取轨道扣件服役状态数据;利用视频线将所述服役状态数据传输至基于深度学习的检测系统;连接控制线路使其达到司控室可以控制启停和接收检测结果;方法包括:获取轨道扣件服役状态数据;将所述运行数据分别输入至多个故障检测模型和一正常检测模型中,所述多个故障检测模型和所述正常检测模型为基于深度学习的自编码器,所述多个故障检测模型分别用于编码得到轨道扣件不同故障种类和不同故障种类组合的特征,所述正常检测模型用于编码得到轨道扣件正常的特征;如果所述正常检测模型的输岀特征与正常特征库中的特征不匹配,将所述多个故障检测模型输岀的特征分别与对应故障种类或组合的特征库进行比对;根据比对结果识别轨道扣件的故障。本实施例員实现轨道扣件故障的自动化识别,提高识别精度。


    Access

    Download


    Export, share and cite



    Title :

    Vehicle-mounted rail fastener service state detector based on deep learning


    Additional title:

    一种基于深度学习的车载式轨道扣件服役状态检测器


    Contributors:
    XU QI (author) / YANG NENGPU (author) / MIAO NING (author) / NIE RENJIA (author) / ZHENG WEI (author) / LONG LIXING (author) / SHI PENGHUI (author) / WU YONGGANG (author) / TANG CHUAN (author) / WEI YANPING (author)

    Publication date :

    2023-03-24


    Type of media :

    Patent


    Type of material :

    Electronic Resource


    Language :

    Chinese


    Classification :

    IPC:    G06V / B61K Andere Hilfseinrichtungen für Eisenbahnen , OTHER AUXILIARY EQUIPMENT FOR RAILWAYS / G06N COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS , Rechnersysteme, basierend auf spezifischen Rechenmodellen



    Rail fastener service state intelligent detection robot

    WANG MENGJUN / WANG YUHUA / YANG HONG et al. | European Patent Office | 2024

    Free access

    Rail fastener detection of heavy railway based on deep learning

    Yuan Cao / Zihao Chen / Tao Wen et al. | DOAJ | 2023

    Free access

    Rail fastener

    CORONEL WOLFGANG E | European Patent Office | 2018

    Free access

    RAIL FASTENER

    CORONEL WOLFGANG E | European Patent Office | 2018

    Free access

    Rail transit vehicle-mounted state detection device

    GUO CHUNYAN | European Patent Office | 2020

    Free access