基于模型配准的钢轨动态轮廓匹配方法

Rail dynamic contour matching method based on model registration

在钢轨日常维护过程中,会面临运行过程中车体各种振动对钢轨轮廓检测系统造成的影响,消除这种影响并获取准确的钢轨动态轮廓数据是必须要解决的问题。在传统最近点迭代算法的基础上,根据振动影响下的测量变形轮廓与标准轮廓存在仿射映射的关系,提出了基于提取钢轨轮廓曲线特征区域完成最佳拟合的模型配准算法,完成了钢轨轮廓的动态匹配,最后通过试验模拟了不同情况下的振动来验证本方法的可行性。试验结果表明,经过动态轮廓匹配后在垂直磨耗和侧面磨耗处的标准差分别为0.002 76、0.002 643mm,重复性精度均能达到0.005mm,本方法基本上能消除振动对测量数据的影响,提高了测量精度。

In the course of routine maintenance of rail, we will face the impact during operation of the rail system, causing the body contour detection of a variety of vibration, eliminating this effect and obtaining accurate rail dynamic profile data is a problem that must be addressed. Based on the traditional characteristics recent point iterative algorithm, in accordance with a standard profile the profile deformation measured under the influence of vibration presence affine mapping model is proposed to complete the registration algorithm based on a best fit curve outline feature region extraction rail, complete dynamic matching profile rail, and finally to verify the feasibility of the method by vibration simulation experiments under different conditions. Experimental results show that after dynamic contour matching, the standard deviations at vertical wear and side wear are 0. 002 76 mm and 0. 002 643 mm respectively, repeatability can reach 0. 005mm, the present method substantially eliminates the influence of vibration on measurement data, improve the measurement accuracy.