Font Size: a A A

Research On Image Recognition Method Of Freight Car Bogie Based On Shape Template

Posted on:2022-05-23Degree:MasterType:Thesis
Country:ChinaCandidate:L LeiFull Text:PDF
GTID:2492306740959449Subject:Traffic and Transportation Engineering
Abstract/Summary:PDF Full Text Request
With the rapid progress of the national economy and the continuous advancement of the reform of railway transportation,the application of railway vehicles has been transformed to the mode of heavy load,high speed and high density.The high efficiency and accuracy of train fault detection are of more prominent significance to the safe operation of trains.In this thesis,the bogie images collected by trackside equipment are taken as the research object to carry out the research on the automatic positioning and fault identification of bogie components.This thesis mainly studies the following aspects:(1)In view of the complex types of bogie components,the different fault characteristics and the demand for real-time detection,this thesis proposes to take the method of locating different components first,and then putting forward the fault identification algorithms according to the different fault characteristics of different components.Aiming at the image quality and texture information are easily affected by different lighting conditions,different types of bogies and obstructions.A component location method based on contour features is proposed,and the automatic location of different components was completed by using the contour features and the established geometric mathematical model.Through component positioning of real image,it is found that contour feature based positioning algorithm is robust to surface texture changes in the bogie image.(2)In this thesis,the angle feature existing in the image of the deflection fault of the locking plate are analyzed,and proposes two methods for the identification of the deflection fault of the locking plate based on the shape template and the truncation feature.By comparing the detection results of real and simulated images,the shape template-based identification algorithm can accurately identify the deflection fault of locking plate at different angles(deflection angle > 5?)and the identification rate can meet the requirements of real-time detection.At the same time,it is found that the angle feature of the lower edge of the locking plate is taken as the basis to judge whether the locking plate is deflected,which has high reliability.(3)In this thesis,two fault identification methods for side frame key loss based on template matching and shape template are proposed.By comparing the size of input image and number of templates in template set influence on the recognition speed based on template matching algorithm and the detection results of two identification algorithms on real image,it is found that the identification algorithm based on template matching detection rate is higher,but the identification speed is easily affected by the template number of template set and the size of the input image,the influence of and interception greatly influenced by the size of templates and the size of the input image,when the size of the image to be detected increases to a certain size,it can not meet the requirements of real-time detection.The identification algorithm based on shape template can not only achieve a higher detection rate,but also achieve a higher identification rate,which can meet the real-time requirements of fault detection.
Keywords/Search Tags:image processing, bogie, fault detection, components localization, shape template, locking plate, side frame key
PDF Full Text Request
Related items