摘要
双块式轨枕是我国高速铁路建设采用最多的预制构件,传统人工检测方法效率低、精度差,无法完全满足检测需求。为解决该问题,基于高精度三维激光扫描仪和高分辨率工业相机,研发一种双块式轨枕外形外观质量自动化检测系统,该系统将检测系统与生产线进行集成,使检测系统具备自动化在线检测功能。另外,研究轨枕三维点云和高清图像自动化采集技术、三维点云数据分析处理技术、轨枕图像分析处理技术,通过点云分割、模型重建算法以获得轨枕关键部件的模型数据,实现轨枕外形尺寸自动检测;通过图像模板匹配和图像深度学习算法,实现轨枕编号与轨枕裂纹智能识别。研究结果表明,系统的检测精度可达到0.1 mm,检测效率优于3 min/组,裂纹识别准确率优于95%,检测项目和检测精度满足相关规范要求,并具有自动化、实时性、高精度、高效率的优点。
The double block concrete sleeper is the most used prefabricated components in Chinese high-speed railway construction.At present,it is mainly produced by assembly line prefabrication.The traditional manual inspection method cannot fully satisfy the requirements of the double block sleeper.In order to solve the problem of high accuracy requirements and heavy inspection tasks,this paper proposed a double block sleeper shape and appearance quality automatic inspection system based on high-precision 3D laser scanner and highresolution industrial camera.The inspection system was integrated with production line,so that the inspection system was automated operated online with inspection function.This paper researched the automatic acquisition technology of 3D point cloud and high-definition image of double block sleeper,3D point cloud analysis and processing technology of double block sleeper,images analysis and processing technology of double block sleeper.Through point cloud segmentation and model reconstruction algorithm,the key components model of double block sleeper was obtained,automatic detection of the shape dimensions was realized.Through image template matching and image deep learning algorithm,intelligent recognition of sleeper numbers and cracks was realized.The inspect accuracy of the system can reach 0.1 mm,the detection efficiency is better than 3 minutes/group,the crack recognition accuracy is better than 95%,and the detection items and accuracy meet the relevant specifications.The system has the characteristics of automation,real-time,high precision,and high efficiency,can effectively solve the quality control problems of double block sleeper prefabricated factory.
作者
秦守鹏
QIN Shoupeng(China Railway Design Corporation,Tianjin 300308,China)
出处
《铁道勘察》
2024年第5期114-119,共6页
Railway Investigation and Surveying
基金
中国铁路设计集团有限公司科研计划重点课题(2022A02538010)。
关键词
高速铁路
双块式轨枕
三维激光扫描
工业相机
外形质量
裂纹检测
high-speed railway
double block sleeper
three-dimensional laser scanning
industrial camera
shape quality
crack detection
