期刊文献+

针对低像质轨道车辆车顶螺栓松动检测方法研究

Research on the Detection Method of Loose Roof Bolts of Low-quality Rail Vehicles
原文传递
导出
摘要 在轨道车辆车顶螺栓松动检测中,由于维护厂棚内列车的上方架设有高压供电线路,导致对车顶器件的螺栓松动情况进行扫描式视觉检测的组件需要架设在更高的位置,使得相机采集的螺栓图像像素数量不多,螺栓图像质量较低,给螺栓松动的判断带来困难的问题。提出了一种针对低像质轨道车辆车顶螺栓松动检测方法,实现了车顶区域螺栓松动的智能检测。检测时首先对线阵相机采集后的轨道车辆车顶图像进行拼接,并定位车顶螺栓的安装区域;其次将该区域细分为出风口、盖板和风机等几个子区域并导入基于深度学习的目标检测方法,从而确定分区内的螺栓位置;然后通过定位后的螺栓中心点坐标绘制等距半径圆,提取螺栓上的防松线并进行骨架提取与直线拟合;最后,通过计算两条直线之间的角度判断螺栓是否松动。实验结果表明,此方法的轨道车辆车顶螺栓松动的旋转角度,最大相对误差为5.8%,并且螺栓松动检出率达到80%以上,具有一定的工程实用价值。 In the detection of loose roof bolts of rail vehicles,due to the high-voltage power supply line on the upper frame of the train in the maintenance shed,the components for scanning visual inspection of the loose bolts of the roof devices need to be erected in a higher position,so that the number of pixels of the bolt image collected by the camera is limited,and the quality of the bolt image is low,which brings difficult problems to the judgment of loose bolts.Therefore,this paper proposes a method for detecting loose roof bolts of low-quality rail vehicles,which realizes the intelligent detection of loose bolts in the roof area.During the inspection,the roof image of the rail vehicle collected by the line scan camera is first stitched,and the installation area of the roof bolt is located.Secondly,the area is subdivided into several sub-areas such as air outlets,cover plates and fans,and the target detection method based on deep learning is introduced to determine the bolt position in the partition.Then,the equidistant radius circle is drawn by the coordinates of the center point of the bolt after positioning,the anti-loosening line on the bolt is extracted,and the skeleton extraction is carried out and fitted to the straight line.Finally,whether the bolt is loose is judged by calculating the angle between the two straight lines.The experimental results show that the maximum relative error of the loose rotation angle of the roof bolt of the rail vehicle is 5.8%,and the detection rate of the loose bolt reaches more than 80%,which has certain engineering practical value.
作者 王正家 谷峰 曾臻 WANG Zhengjia;GU Feng;ZENG Zhen(School of Mechanical Engineering,Hubei University of Technology,Wuhan 430068,China;Hubei Key Lab of Modern Manufacture Quality Engineering,Wuhan 430068,China)
出处 《机械设计与研究》 CSCD 北大核心 2024年第2期220-224,共5页 Machine Design And Research
基金 国家自然科学基金资助项目(51275158)。
关键词 轨道车辆 螺栓松动 目标检测 深度学习 HALCON railway vehicles loose bolts object detection deep learning Halcon
  • 相关文献

参考文献11

二级参考文献88

共引文献61

相关作者

内容加载中请稍等...

相关机构

内容加载中请稍等...

相关主题

内容加载中请稍等...

浏览历史

内容加载中请稍等...
;
使用帮助 返回顶部