作者：陈仕明¹, 杜馨瑜², 孙淑杰², 赵鑫欣²

作者单位：1. 中国铁道科学研究院研究生部，北京 100081;
2. 中国铁道科学研究院集团有限公司基础设施检测研究所，北京 100081

关键词：限界系统;振动补偿;计算机视觉;动态测量

摘要：

针对列车车体的随机振动影响到限界系统的测量结果的问题，提出一种基于机器视觉的振动补偿测试系统。该系统以轨道为参考建立轨面基准坐标系，结合车体的振动特性，根据激光摄像组件所测位移，推导出车体振动偏移补偿计算方法，用于补偿激光扫描传感器的测量结果。为测试系统系统动态精度，提出基于扩展卡尔曼滤波的多传感器信息融合算法。动静态试验结果表明：实测误差接近于理论分析误差，验证了算法的可行性。对于当前的传感器安装高度，补偿精度可达到1 mm。当前已完成线路上的动态测试，将应用于综合巡检车中。

Research on vibration compensation for clearance system based on machine vision
CHEN Shiming¹, DU Xinyu², SUN Shujie², ZHAO Xinxin²
1. Graduate Department of China Academy of Railway Sciences, Beijing 10081, China;
2. Infrastructure Inspection Research Institute, China Academy of Railway Sciences Corporation Limited, Beijing 100081, China
Abstract: In order to overcome the negative influence on the measurement accuracy of the clearance system due to train vibration, a vibration compensation algorithm based on machine vision is proposed. In the system, the world coordinate system was constructed with tracks as the reference system. The vibration compensation algorithm is derived according to displacement measured by Camera&Laser and combining vibration characteristics, which is applied to compensate the results of laser scanning sensor. The EKF algorithm is proposed for testing system dynamic precision. The results show that the measured error is close to the theoretical analysis error, which verifies the feasibility of the algorithm, and the accuracy reaches 1 mm. Now the system is ready to be installed in the comprehensive inspection train.
Keywords: clearance system;vibration compensation;computer vision;dynamic measuring
2022, 48(4):129-135  收稿日期: 2021-04-06;收到修改稿日期: 2021-06-25
基金项目: 中国国家铁路集团有限公司科技研究开发计划课题(2017G003-l)
作者简介: 陈仕明（1995-），男，江苏盐城市人，博士研究生，专业方向为轨道几何动态测量与基础设施检测
参考文献
[1] 孙淑杰, 王凡, 赵鑫欣, 等. 铁路建筑限界检测系统及应用[J]. 中国铁路, 2017(10): 82-86
[2] 占栋, 于龙, 邱存勇, 等. 钢轨轮廓测量中的车体振动补偿问题研究[J]. 仪器仪表学报, 2013, 34(7): 1625-1633
[3] 陈唐龙, 于涤, 陈耀坤. 接触网检测车振动补偿研究[J]. 西南交通大学学报, 1999, 34(4): 461-465
[4] 王铁成, 李玉龙, 徐练, 等. 轨道车辆动态限界测试系统设计[J]. 中国测试, 2020, 46(3): 113-117
[5] 刘涛, 吴积钦, 曾明. 非接触式接触网检测车车体振动位移补偿系统[J]. 电气化铁道, 2007(5): 21-24
[6] 赵鑫欣, 王胜春, 王昊, 等. 车载动态条件下钢轨廓形快速追踪算法[J]. 中国铁道科学, 2020, 41(6): 145-155
[7] 占栋, 于龙, 肖建. 基于计算机视觉的接触轨检测车振动补偿方法及应用[J]. 铁道学报, 2013, 35(1): 31-36
[8] 张冬凯, 高仕斌, 于龙, 等. 车体振动对接触网检测的影响分析及补偿方法研究[J]. 铁道学报, 2019, 41(9): 43-50
[9] 薛鹏, 柴晓冬, 郑树彬. 基于视觉的车辆与轨道相对振动状态测量方法研究[J]. 计算机测量与控制, 2016, 24(1): 39-41
[10] 占栋, 于龙, 肖建, 等. 轨道检测中激光摄像式传感器标定方法研究[J]. 机械工程学报, 2013, 49(16): 39-47
[11] XU Q, CHAI X, ZHENG S, et al. Template matching for profile measurement based on Levenberg-Marquardt algorithm[J]. American Journal of Industrial & Business Management, 2014, 4(7): 370-375
[12] TEDALDI D, PRETTO A, MENEGATTI E. A robust and easy to implement method for IMU calibration without external equipments[C]//2014 IEEE International Conference on Robotics and Automation (ICRA). IEEE, 2014: 3042-3049.
[13] TRAWNY N, ROUMELIOTISS I. Indirect Kalman filter for 3D attitude estimation [R]. Twin Cities: University of Minnesota, Department of Computer Science and Engineering. 2005: 2-23.
[14] QIN L, DA F P. Calibrating lens distortion of camera in all directions[J]. Chinese Journal of Scientific Instrument, 2010, 31(9): 2022-2027