Study On Geometric Parameters Measurement Technology Of Wheel Tread Based On Line Structured Light

The wheelset,as the key part of the train running,is very crucial for the safe operation of the train.The working environment of wheelset,which bears the static and dynamic load of train,is bad.The contact with rail and brake causes friction.Meanwhile,wheelset is corroded frequently by rainwater,oil and chemical agent.Inevitablely,wears and defects appear after a period.Thus,it is necessary to detect the relevant geometric parameters to ensure the safety of running.At present,the method of manual measurement is widely used in our country,which has high technical requirements for maintenance personnel and low measuring efficiency.Therefore,it is of great significance to study the automatic measurement method of wheelset geometric parameters.Because of its high accuracy and rapidity,the method of line structured light non-contact measurement has been used widely.In this paper,on the basis of the theory of image processing,the algorithm of automatic measurement of wheelset geometric parameters is studied based on line structured light method.The main work and innovations in the paper are as follows:Firstly,the current measuring methods are compared,and the question that the laser plane can't pass the axis in the actual measurement are analyzed.Thus,the string-aided positioning scheme based on the vertical diameter theorem is proposed,which makes full use of the advantages of non-contact measurement technology of line structured light and the traditional static measurement.The above two methods are effectively integrated for the location of base point.Meanwhile,the device which can preliminarily meet the requirements is designed,and the measurement of key parameters,including rolling diameter,flange thickness and flange height are implemented by accurate positioning.Secondly,the key image processing technologies are studied in the process of automatic measurement algorithm in this paper.Various threshold segmentation algorithms are analyzed and experimented.Subsequently,the optimal algorithm can be selected reasonably so that the threshold value is obtained automatically.Meanwhile,the skeleton extraction algorithm is also optimized.The pixels meeting the conditions of deletion in Zhang thinning algorithm are classified according to their cases of eight neighborhood.LUT,used to improve the efficiency of skeleton extraction,is made.Then,the limitation of least square method is discussed.Aiming to enhance the robustness of the measurement algorithm,instead of least square method,RANSAC algorithm is used to fit the flange inside line equation.Through designing different simulation experiments and evaluation indexes,the fitting results are analyzed.The conclusion shows that the RANSAC algorithm,which has strong robustness under the experimental condition,is better than the least square method and can satisify the actual needs.Finally,the measurement of geometric parameters of wheelset is implemented by programming based on the platform of MFC.Through field experiments,the manual measurement results and the system measurement results are compared and analyzed.The feasibility of the geometric parameter automatic measurement algorithm are proved.