Study On Spot Central Positioning Method Of Railway Subgrade Surface Settlement Monitoring System

The ballastless track has become the main structural type of high-speed railway.The ballastless track has good track regularity and its track geometry can keep for a long time.With the rapid development of high-speed railways,the requirements for subgrade surface settlement monitoring are also increasing.The traditional manual inspection method can't meet the needs of current operational management conditions.With the development of signal processing theory and related technologies such as electronics and computers,machine vision has developed rapidly.Based on the machine vision,the railway subgrade surface settlement monitoring system obtains the position data difference before and after the settlement of the target by positioning the laser image center to reflect the settlement deformation of the subgrade at the monitoring point.Target image extraction is performed based on the spot image characteristics.The sub-pixel level centroid is obtained by the local gray scale centroid method.Using the monocular vision system combined with the position information of the characteristic points of the target surface,the system can solve camera parameters in real time,and the centroid localization result is corrected.The main research contents are as follows.Firstly,the methods and data for machine vision inspection and spot central positioning method of railway subgrade surface settlement are collected and investigated.A mathematical model of the system spot centering technology is established.Based on the principle of camera imaging,a camera model of the monitoring system is established.The measurement principle of the relative position of the target surface and the camera in the system is obtained.Using Zhang's method to reduce the difficulty of camera calibration,complete the calibration of the laboratory camera parameters.Secondly,the image background interference and other factors,which are common in actual measurement,cause the blurring of the edge of the spot image.Based on the image using adaptive median filtering to remove image noise,the maximum between-cluster variance threshold is used to segment the cross structure of the spot image.Then,the morphological opening and closing operation is used to smooth the cross boundary,and finally the spot image with research value is obtained.Thirdly,it is pointed out that the traditional spot center positioning technology will cause a large error in the positioning of the laser image when the distance between the laser and the target surface is far.A two-step precise localization algorithm based on pixel gray value in local region is proposed,which is also called local center of gravity method.Then,in view of the environmental factors such as vibration during the railway operation,the relative deflection of the target surface and the camera will affect the positioning result of the spot center.A monitoring target pose measurement system is designed.Based on the monocular vision,the system can solve camera parameters in real time,and the correction of the spot center positioning result is realized.Finally,an experimental platform for spot central positioning method of railway subgrade surface settlement monitoring system is built.The experimental results show that the maximum error of settlement value is 0.16 mm,the minimum error is 0.1mm,and the total experimental time is about 1.5s.The spot centering positioning method has higher calculation accuracy and shorter time,which is suitable for the flawless track surface settlement monitoring system.