Research On Positioning And Classification Method Of Metro Rail Fastener Image

Urban rail transit,as an economic and high-volume vehicle,plays a very important role in the transportation industry of all countries in the world.With the rapid expansion of China’s urban rail transit in recent years,the importance of rail safety is becoming increasingly prominent,A good safety state of urban track cannot be separated from the state of fasteners,which is an important connecting part.Therefore,it is very important to locate,detect and classify fasteners on urban track.Nowadays,the huge modern subway network constantly puts forward new goals and requirements for the safety inspection of urban rail transit,and it is necessary to conduct longterm and effective inspection and maintenance of lines to ensure the healthy operation of urban rail transit.For these heavy work and objective needs,an efficient,scientific and automated detection method is urgently needed to ensure the safety and reliability of urban rail transit.The main research content of this paper is to put forward a method of coupler positioning and classification detection according to the development status of urban rail transit in recent years.The specific work is as follows:Firstly,the track image acquisition and the preprocessing of the original image are studied.The first step is to collect the subway track image.The subway track detection equipment used in this paper is a new type of track inspection instrument.It takes into account the unique operation conditions of subway rail transit and adopts a new design.It can select appropriate components for detection according to the actual situation of different subway lines.The algorithm used in the preprocessing part is to grayscale the track image,then smooth the image and Gaussian Image Denoising.The second chapter of this paper can reduce the amount of image data,improve the quality of the image,for the subsequent positioning and classification work to reduce the difficulty.Coupler positioning is an important step in the process of detecting railway coupler integrity by using image automatic recognition technology.In this paper,a new method of coupler positioning is put forward based on the characteristics of the integrated bed and the application of the new inspection instrument in the integrated bed.Firstly,the image was compressed and then the Sobel edge was detected.Then the gray level of the processed image was projected to the Y-axis to determine the rail position.Finally,the rail is used as the vertical coordinate axis,and the coupler is located and extracted according to the rule of coupler installation interval.The experimental results show that this method can filter the noise in the image well and extract the fasteners on the subway bed effectively.In order to accurately describe the key information of the image and distinguish the extracted fastener image,in Chapter four,the feature value extraction of fastener image based on gray level co-occurrence matrix is carried out.Four eigenvalues of ASM,ENT,CON and COR are extracted,and the extracted eigenvalues are processed and analyzed.In this chapter,the gray level co-occurrence matrix method is firstly applied to the environment of urban rail integrated bed.The image information is simplified and the eigenvalue data that can accurately describe the fastener image information is obtained,which lays a foundation for the subsequent work of fastener recognition in this paper.In the fifth chapter,BP neural network is used to train and classify the four eigenvalues.The experimental results show that a series of processing for the track image of the monolithic track bed in this paper can achieve the classification of fastener state,and has a high accuracy.