基于图像的轮对踏面轮廓测量方法

The wheel tread profile measurement method based on image processing

轮对作为动车运行的重要部件,如果出现故障,将引发不可预想的后果,因此如何实时地、有效地、精确地检测车轮在列车众多安全问题中显得尤其重要。文中在使用非接触式检测车轮,通过CCD采集系统获得原始的轮对踏面曲线的基础上,通过图像降噪处理、曲线提取和曲线拼接一系列方法,实现获取车轮踏面完整轮廓。文中在曲线提取和拼接的算法部分,通过对比已有的方案和方法,提出了适合轮对踏面轮廓的改进算法并通过实验进行分析和验证。通过实验证明,文中提出的车轮踏面轮廓测量方法易于实现、适用范围广。

Wheel as an important part of the running of the bullet train,if there is a failure,it will lead to unexpected consequences,therefore,it is especially important that how to detect the wheel in real time,effectively and accurately in many safety problems of trains. On the basis of the use of non-contact detection of the wheel,with the CCD acquisition system to obtain the original wheel tread surface curve,through a series of methods for image noise reduction,curve extraction and curve matching,this paper acquires the complete contour of the wheel tread. In the part of the algorithm of curve extraction and mosaic,by comparing the existing schemes and methods,it presents an improved algorithm for wheel tread profile and analyzed and validated by the experiments. In addition,the experimental results show that this method is easy to implement and can be applied widely to the measurement method of wheel tread profile.