| Our country’s railway industry has developed rapidly in the past two decades.As of the end of the first quarter of 2021,the national railway operating mileage has reached 146,700 kilometers.The rail is an important infrastructure that guides the train forward and supports the operation of the train.The change in its outline size directly reflects the quality of the rail in service.When the outline size is severely deteriorated,it will cause safety hazards to the running train or even cause serious traffic accidents.In order to ensure the safe operation of trains,it is necessary to quickly and efficiently detect changes in rail profile dimensions.At present,the rail contour detection equipment based on computer image processing technology is mainly used in the field of rail contour detection in our country,which plays a certain role in preventing train accidents caused by the change of rail contour size,but the detection equipment has the problem of insufficient detection efficiency.Aiming at this problem,this paper designs a high-speed track inspection system based on structured light vision measurement technology and FPGA hardware acceleration technology on the basis of existing research results.In this paper,the basic structure and working principle of the detection system are explained in more detail.The actual imaging model of the camera is derived based on the actual working conditions,and the linear least squares method is used according to the characteristic point image coordinates and world coordinates obtained by the needle-shaped target calibration method.The precise calibration parameters of the camera have been obtained.Secondly,it analyzes the image preprocessing algorithm and the light strip center extraction algorithm of various precision levels in detail,and combines various hardware acceleration optimization strategies to achieve three different precision levels of extreme value method,gray barycentric method and Hessian matrix method.Hardware acceleration of the light bar center extraction algorithm.The experimental results show that the processing efficiency of the image algorithm through FPGA hardware acceleration is greatly improved compared with the traditional industrial computer processing.Then,by analyzing the influence of various random vibrations on the alignment results of the rail profile under dynamic measurement conditions,a dynamic measurement alignment error compensation model based on the feature points of the gauge as the compensation reference is established.By calculating the translational and rotational components between the characteristic points of the dynamic and static gauges,and substituting them into the dynamic measurement alignment error compensation model,the purpose of reducing the rail profile detection error is achieved.Finally,considering the actual working conditions,FPGA is selected as the core hardware,and the hardware design of the rail profile measurement system based on FPGA is carried out.At the same time,the software design of the detection system is carried out under the Windows development environment.By comparing and analyzing the image data and the result coordinate data through field statistics,the processing efficiency of the system is verified.The statistical results show that the processing efficiency of the system is increased from the original 60 frames/sec to 180 frames/sec,and the matching train detection speed is increased from the original 54km/h to 160km/h. |
