Design Of Visual Speed Measurement System For Urban Rail Train

In the train operation control system,the speed information is of great significance to ensure the driving safety and maintain the reliable operation of the train.With the rapid development of artificial intelligence technology,target detection and recognition technology based on computer vision is applied more and more in engineering.Therefore,developing a set of urban rail train speed measurement system based on computer vision has important engineering application prospects.Based on the analysis of traditional speed measurement methods,this thesis develops and designs a set of real-time speed measurement system based on computer vision technology,which is low-cost,high reliability,easy to install and disassemble,independent of other auxiliary equipment and suitable for urban rail trains.The main contents of this thesis are as follows:(1)This thesis analyzes the research status of the mainstream speed measurement methods of urban rail trains at home and abroad,summarizes the advantages and disadvantages of each speed measurement method in application,analyzes the application prospect of image processing technology in computer vision in train speed measurement,and puts forward to explore the speed measurement of Urban Rail Trains Based on computer vision technology.(2)This thesis puts forward the design goal of urban rail train visual speed measurement system,and introduces the visual speed measurement principle of the speed measurement system,as well as the overall structure and software module composition of the system.(3)According to the functional requirements and component selection,the mechanical structure is designed,mainly including waterproof structure,camera fixing structure and vibration damping structure.The selection of light source,camera and lens is completed through functional analysis.Considering the heating characteristics of the equipment,the heat dissipation analysis of the structure is carried out.The vibration simulation analysis of the mechanical structure is carried out with ADAMS software,and the optimal parameters of vibration isolation are studied.(4)This thesis analyzes the feasibility of the camera’s on-off output signal(DO)as the LED light source control signal,designs the control-drive integrated circuit board which integrates the control and drive functions and can output PWM signal,and draws the circuit schematic diagram of the control-drive integrated circuit board,including voltage conversion module,communication module,control module and drive module,Finally,the corresponding relationship between the camera do output signal frequency and the PWM signal of the control-drive integrated circuit board is given.(5)The image acquisition and preprocessing are completed,including image noise reduction and distortion correction.This thesis analyzes and selects the incremental PID control algorithm for the adaptive lighting of LED light source,introduces the principle and application characteristics of common occlusion detection algorithms,and then designs the camera occlusion discrimination algorithm suitable for the speed measurement system according to the actual application scenario of this design.The performance indexes of mainstream target detection algorithms are analyzed.Combined with the real-time requirements of this design,the improved Yolo algorithm is selected to detect sleepers.Then the program design of each algorithm module is completed,and the functional experimental verification is carried out.(6)The camera calibration and simulation experiment platform are built,and the speed measurement system software is designed and developed.The accuracy of speed measurement is verified by track simulation software.