轮对轮辋厚度检测中图像配准方法研究

针对轮辋厚度参数的在线检测,提出了基于双相机的轮辋厚度参数检测方法,其中图像配准是影响该参数计算的重要因素。在分析轮对轮廓线图像特点的基础上,提出了利用配准板特征点的图像配准方法,设计了图像配准步骤,采用仿射变换作为图像配准中的变换模型。为保证配准精度,在综合考虑配准板上特征点的数量和特征点的分布的基础上设计了配准板。通过图像配准实验,提取了配准板特征点坐标,求解了仿射变换矩阵。实验表明,该方法的配准精度为2个像素。

Research on image registration method for detecting wheel set rim thickness

The method about detection of rim thickness parameter based on dual camera is proposed. Image registration is an important factor affecting the parameter calculation. Based on analyzing the characteristics of images of wheel contour, an image registration method is proposed, which relies on the feature point on registration plate. The image registration steps are designed. The affine transformation model is determined. The registration plate is designed for the consideration of the distribution and the number of feature points on the registration plate. The accuracy can be guaranteed. Through the experiment of image registration, the coordinates of feature points is extracted. The affine transformation matrix is solved. In the experiment, the accuracy is 2 pixels.