Research On Key Technology Of Urban Track Service State Detection System

Since stepping into the 21st century,the rail transportation industry of China has entered a new stage of leap-forward development.Rail defects and tunnel damage are two important factors that threaten operation comfort,convenience,safety and environmental protection.So strengthening the rail detection is conducive to improve the stability and safety of train's operation,and is also helpful to reduce energy loss so as to achieve green environmental protection.However,the rail inspection and maintenance in China is facing a lot of difficulties and limitations such as short sky window,poor operating environment,low level intelligence equipment,low operation efficiency and so on.Therefore,this thesis is a study on the development of fast and multi-functional rail inspection equipment and rail detection methods of high efficiency and high precision.The main research contents are as follows:At first,the high precise trajectory of the carrier platform is computed by combining LiDAR/IMU/Odometer.With the trajectory and calibration parameters put into multi-source data fusion model the 3D point cloud of track full section can be calculated.Also the high precise rail cross section is calculated with the motion correction model.On these basis,the rail's geometric parameters can be calculated directly from definitions.Then,an algorithm based on adaptive closest point(AICP)is proposed to detect and classify rail surface defects,which improved the registration accuracy between mea-sured rail section and standard rail section model to sub-millimeter.On this basis,the candidate defective points are located and then form defects area by using the K-means algorithm,after then the decision tree classifier is taken to classify the defects.And then,an algorithm based on wavelet transform and iterative closest point(ICP)algorithm is proposed to identify the surface cracks of tunnel lining.The algo-rithm uses the improved motion correction model to calculate the rail cross section,matches the cross section with standard model roughly using calibration parameter and generates the waveform of deviation between the cross section and standard model.By using the wavelet transform the waveform's high frequency part is filtered,so the pure tunnel lining cross section is obtained.On this basis,the ICP algorithm is used to precisely match the pure tunnel lining cross section with the tunnel lining standard model so as to quickly and accurately locate the crack area according to the tunnel lining crack's characteristics and extract the length,width and density of the cracks.At last,on the basis of these studies a set of urban track service state detec-tion system is developed.This system integrates several sensors such as laser scanner,laser profiler,odometer,inertial measurement unit(IMU)and global positioning system(GPS)and obtains 3D point cloud of track full section.The system can achieve the goal of track geometric parameters measurement,rail surface defects detection and tunnel surface crack detection,and the accuracy and performance are verified through a series of field testes.