首页|基于Faster R-CNN的铁路货车心盘螺栓智能检修

基于Faster R-CNN的铁路货车心盘螺栓智能检修

扫码查看
为了改善人工检修心盘螺栓浪费大量人力,检修作业效率过低的缺陷,研究基于Faster R-CNN的铁路货车心盘螺栓智能检修方法.利用机器视觉设备采集铁路货车心盘螺栓图像,对所采集图像进行膨胀与腐蚀处理,获取图像的感兴趣区域.设置图像的感兴趣区域作为Faster R-CNN模型的输入,FasterR-CNN模型利用RPN部分,生成螺栓状态检测区域,通过池化层与全连接层的运算,输出铁路货车心盘螺栓状态检测结果.针对非正常状态的铁路货车心盘螺栓,利用全自动拧紧机通过自动卡紧、自动对位拧紧等操作,完成心盘螺栓的全自动维修.实验结果表明,该方法可以精准检测铁路货车心盘螺栓的正常、松动、丢失等状态,全自动维修非正常状态的心盘螺栓,提升铁路货车心盘螺栓检修的作业效率.
Intelligent maintenance of rail truck tray bolts based on Faster R-CNN
In order to improve the defects of manual maintenance of the tray bolt,which wastes a lot of manpower and has low maintenance operation efficiency,the intelligent maintenance method of railway truck tray bolt based on Faster R-CNN is studied.U-sing machine vision equipment to collect the image of railway freight car core plate bolt,the acquired image is processed by expansion and corrosion,and the interested area of the image is obtained.The region of interest of the image is set as the input of the Faster R-CNN model.The Faster R-CNN model uses the RPN part to generate the bolt status detection region.Through the calculation of the pooling layer and the full connection layer,the bolt status detection results of the railway freight car core plate are output.Aiming at the abnormal state of railway freight car centerplate bolt,automatic tightening machine is used to complete the automatic maintenance of centerplate bolt by automatic clamping and automatic alignment tightening.The experimental results show that the method can ac-curately detect the normal,loose and missing state of the center bolt of railway freight car,automatically repair the abnormal state of the center bolt,and improve the working efficiency of the overhaul of the center bolt of railway car.

faster R-CNNrailway wagonscenter plate boltintelligent maintenancepooling layerfully connected layer

王鹏、孟鑫、赵军、刘宇、杨翠萍

展开 >

国能铁路装备有限责任公司沧州机车车辆维修分公司,河北沧州 061100

中铁工程设计咨询集团有限公司,北京 100055

北京铁科合力科技有限责任公司,北京 100082

Faster R-CNN 铁路货车 心盘螺栓 智能检修 池化层 全连接层

2024

自动化与仪器仪表
重庆工业自动化仪表研究所,重庆市自动化与仪器仪表学会

自动化与仪器仪表

CSTPCD
影响因子:0.327
ISSN:1001-9227
年,卷(期):2024.(5)