Research On The Detection Method Of Wheelset Tread Of Rail Transit Vehicle Based On ICP Algorithm

With the rapid development of urbanization,rail transit accounts for an increasing proportion of urban traffic due to its advantages of not occupying ground space,rapid security,and large capacity.What follows is a significant increase in maintenance for rail transit vehicles.The maintenance of wheel treads is an important part of this.The traditional manual detection method is time-consuming and laborious,and it is prone to misdetection of missed inspections,which brings significant safety risks to rail transportation.Based on this,the automatic detection technology of the wheelset is developed.This paper proposes a wheelset tread surface detection method based on ICP algorithm,which can realize the non-contact measurement of the wheelset.The measurement speed is fast and the operation is simple and easy.High accuracy.Based on the deep analysis of the implementation principle of ICP algorithm,this paper proposes an improved ICP algorithm to complete the wheel tread diameter extraction.The original intention of the algorithm is to register the point cloud model.Because of the inherent defects of the point cloud acquisition equipment,it is impossible to collect the complete point cloud model of the object at one time,and it is necessary to splice the multiple local point clouds collected as a complete point cloud..Using the principle of minimum distance of the corresponding point in the ICP registration algorithm,based on the idea of reverse engineering,a benchmark model with different radii is built to register with the measurement point cloud,and finally the tread diameter of the actual measurement point cloud is obtained.Before obtaining the tread diameter of the wheelset,a point cloud model of the wheel tread is obtained first.The imaging principle of the small hole of the camera is studied and analyzed.We decide to use Zhang Zhengyou calibration method to get the parameters of the CCD.The planar light plane is calibrated using the plane calibration plate,and the centerline coordinates of the line structure are obtained by combining the camera's internal and external parameters.The coordinates of all points obtained by scanning are unified into the same coordinate system,and a complete point cloud of the wheel tread is finally obtained.At the same time,several common CAD model simplification methods are studied.The STL model based on vertex deletion is used to simplify the CAD model.The experimental results show that this method can effectively shorten the time of registration.Finally,the experimental analysis shows the effectiveness of the measurement method.