| 基于Faster R-CNN的铁路扣件定位方法研究 |
| |
| 引用本文： | 白堂博,杨建伟,许贵阳,邱实.基于Faster R-CNN的铁路扣件定位方法研究[J].铁道科学与工程学报,2021,18(2):502-508. |
| |
| 作者姓名： | 白堂博 杨建伟 许贵阳 邱实 |
| |
| 作者单位： | 北京建筑大学机电与车辆工程学院,北京 100044;北京建筑大学城市轨道交通车辆服役性能保障北京市重点实验室,北京 100044;北京建筑大学机电与车辆工程学院,北京 100044;北京建筑大学城市轨道交通车辆服役性能保障北京市重点实验室,北京 100044;北京建筑大学机电与车辆工程学院,北京 100044;北京建筑大学城市轨道交通车辆服役性能保障北京市重点实验室,北京 100044;中南大学土木工程学院,湖南长沙 410075 |
| |
| 基金项目： | 国家自然科学基金资助项目;北京市教委科技计划一般项目 |
| |
| 摘 要： | 准确的扣件定位是进行扣件状态检测、保障轨道交通车辆安全运行的基础,传统的基于图像处理的方法难以满足快速准确智能检测的需要.针对这一情况,提出一种基于改进Faster R-CNN的深度学习理论方法,进行扣件定位.首先,建立检测图像数据集并进行图像标注,然后根据实际扣件图像特点建立Faster R-CNN检测模型,利用标注...
|
| 关 键 词： | 铁路 扣件定位 图像处理 Faster R-CNN |
Research on railway fastener positioning based on Faster R-CNN |
| |
| BAI Tangbo,YANG Jianwei,XU Guiyang,QIU Shi.Research on railway fastener positioning based on Faster R-CNN[J].Journal of Railway Science and Engineering,2021,18(2):502-508. |
| |
| Authors: | BAI Tangbo YANG Jianwei XU Guiyang QIU Shi |
| |
| Institution: | (School of Mechanical-electronic and Vehicle Engineering,Beijing University of Civil Engineering and Architecture,Beijing 100044,China;Beijing Key Laboratory of Performance Guarantee on Urban Rail Transit Vehicles,Beijing University of Civil Engineering and Architecture,Beijing 100044,China;School of Civil Engineering,Central South University,Changsha 410075,China) |
| |
| Abstract: | Accurate fastener positioning is the basis for fastener condition detection and the ensurance for railway vehicles safety.The traditional method based on image processing is difficult to meet the need of rapid and accurate intelligent detection.In view of this,a deep learning method based on improved Faster R-CNN was proposed in this paper to locate fasteners.Firstly,the detection image data set was established and labeled.Then,the Faster R-CNN detection model was built according to the characteristics of the fasteners image.The proposal region border information in the model was optimized by using the labeled data to improve the recall efficiency and precision.Through the analysis and verification of actual detection data,the method proposed in this paper improves the detection efficiency and positioning effect.Compared with other target detection methods,it has higher recall rate and identification accuracy,and can locate fasteners quickly and accurately. |
| |
| Keywords: | railway fastener positioning image processing Faster R-CNN |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
|
