摘要
在货车运行中,螺栓丢失是严重危及行车安全的故障类型,在TFDS实际检测作业中,采集的货车关键部位图像主要由人工判别。货车运行故障动态图像检测系统(TFDS)是我国铁路车辆运行安全监控系统的重要组成部分,当前TFDS系统人工检车的工作量大、检车时间短,列检质量容易受列检人员影响。该文将深度学习算法和故障图像智能识别技术引入货车检修作业,通过对货车螺栓的检测,实现了对货车螺栓丢失故障的智能诊断。实验结果表明,提出的算法对螺栓的识别率达到了96%以上,自动识别算法能够减少人的主观性对检修作业的影响,提高检修作业质量和作业效率,提高TFDS系统的安全保障能力。
Trouble of moving freight car detection system(TFDS)is an important part of railway traffic security safeguard equipment monitoring and management system.The key parts of freight car are mainly discriminated by manual from collected images.At present,the manual inspection of TFDS system has a large workload and a short inspection time,and the quality of the list inspection is easily affected by the list inspection personnel.Aiming at the shortcomings of the current TFDS system in practical application,by introducing deep learning,an intelligent recognition algorithm for freight car bolts based on deep learning is studied,so as to realize the recognition of railway freight car bolts.The results show that the recognition rate of the proposed algorithm for bolts reaches more than 96%.The automatic recognition algorithm can reduce the influence of human subjectivity on the overhaul operation,improve the quality and efficiency of the overhaul operation,and improve the safety guarantee ability of the TFDS system.
作者
何宇强
肖致明
张韦昱
邱霁
HE Yuqiang;XIAO Zhiming;ZHANG Weiyu;QIU Ji(Guoneng Shuohuang Railway Development Co.,Ltd.,Cangzhou 062350,China;School of Mechanical,Electronic and Control Engineering,Beijing Jiaotong University,Beijing 100044,China)
出处
《自动化与仪表》
2025年第4期104-108,共5页
Automation & Instrumentation
基金
国家能源集团科技创新项目(GJNY-21-65)
北京交通大学人才基金项目(2024XKRC042)。
