Railway vehicle wheel profile parameters measurement under incomplete curves conditions

The measuring of wheel set geometrical profile is important for the train vehicle running safety. Using light shot digital image processing method to detect wheel profile parameters is becoming the mainstream technology. The precision of measurement, however, is influenced by the sand sprinklers on bogies. The laser lights would be overshadowed by sand sprinklers, which will influence the classification of laser light curves and then the calculation of wheel profile parameters. These influences could not solve by traditional image classification methods. This paper gives a solution for this influence from camera calibration model aspect which solves this problem by project light curves from image to 3-D calibration model, and the solution has been verified by experiment. The experimental results show that the methods can solve the light curve classification problem and ensure the precisions of flange height and flange thick within 0.2mm and wheel diameter within 0.6mm.