科技创新与应用2026，Vol.16Issue(23)：17-19,25,4.DOI:10.19981/j.CN23-1581/G3.2026.23.004
基于机器视觉的轨道爬行位移智能化测量应用研究
摘要
Abstract
Rail creep is a longitudinal displacement phenomenon of railway tracks caused by temperature variations,train starting/braking,and subgrade deformation.The issues it triggers account for about one-third of all railway defects,seriously threatening operational safety.Existing measurement methods(such as the string line method and optical instrument method,et al)suffer from low accuracy and inefficiency,while contact sensors(such as displacement sensors and fiber optic sensing,et al),though more accurate,involve complex installation and maintenance and can compromise operational safety.To address these issues,this study proposes an intelligent measurement system for rail creep displacement based on machine vision.The system can be attached to the mainstream track inspection trolleys used by permanent way departments through a simple mechanical design,significantly reducing equipment costs.By integrating instrument design combinations,it enables vision-based mobile measurement of rail creep,with an operational efficiency of approximately 4-5 km/h.Field test results demonstrate that its accuracy is within±1 mm,meeting monitoring requirements.As a universal system,it can cater to the needs of both high-speed and conventional railways.关键词
轨道爬行/机器视觉/深度学习/图像识别/智能化测量Key words
rail creep/machine vision/deep learning/image recognition/intelligent measurement分类
交通工程引用本文复制引用
刘青松,周新文,罗文彬,张亨..基于机器视觉的轨道爬行位移智能化测量应用研究[J].科技创新与应用,2026,16(23):17-19,25,4.基金项目
中国铁路南宁局集团有限公司科技研究开发计划项目(经24-1) （经24-1）
