Research On Rail Profile Measurement Technology Based On Laser Sensor

Rail transport is a way of efficient and cost-effective land transport.Guided by two parallel rails,it has the advantages of low overall energy consumption,fast operation speed,large capacity,low climate impact,punctuality and continuity.At present,China's railways have entered a stage of rapid development.While boosting economic development and facilitating people's travel,they also attach more importance to the quality and safety of rails.The rail processing units usually use special calipers to measure the rail outline dimensions.This traditional contact measurement has problems such as low measurement accuracy,low efficiency,lack of reliability,high labor intensity,and difficult data storage.A rail profile detection technology based on a point laser sensor performs high-precision and intelligent online detection of a rail profile of a measured rail.According to the demand analysis,this paper selects a static measurement method for the circumferential scanning of the ring rail,which uses a motor to drive a point laser sensor along the ring guide,according to the three measured geometric dimensions of the rail head width,the rail bottom width and the rail height of the rail profile.A 360-degree data acquisition was performed;a corresponding system structure was designed,and modules such as mechanics,electronics,and software were discussed.At the same time,the detection flow of the system was also given.This article focuses on the rail data processing capabilities.According to the steps of data preprocessing,coordinate transformation,fitting matching and error judgment,the characteristics of various algorithms involved are compared and analyzed,and a linear regression model of machine learning is established.Taking into account the actual application environment,the ridge regression algorithm is applied to fit the rail line at the bottom of the rail,and the six points of the rail head,the rail waist,and the rail bottom are selected to carry out rotation and translation matching,thereby converting the collected raw coordinate data points to With the new coordinate system centered on the measured rail,the rail head width,rail height,and rail bottom width of the rail profile are then calculated to obtain higher-accuracy parameter data.In addition,this article describes the system uses Python language to write the user interface for real-time display of the results of its image and detection,the interface is simple and easy to operate.Finally,the error sources of the detection system are analyzed and its uncertainty is deduced.With multiple measurements of 60kg/m selfmanufactured steel rails,the repeatability accuracy of this system is calculated to be between 10?m and 20?m,and the reliability accuracy is within 10?m to achieve the desired target.This can satisfy the relevant production departments and scientific research departments for rail profile detection requirements.