Research On Identification And Application Of On-board Driver Machine Interface Information In Communication Based Train Control System

With the rapid development of urban rail transit in China,communication based train control（CBTC）system has been widely used.During the development process of CBTC system and before it is put into operation,the system function verification is a crucial link.The complete system function verification process includes compiling test cases,testing functions in simulated environment,analyzing and giving feedback on failures in the test,regression testing after software and data version upgrades.It can verify whether the designed system meets the system requirements specification and whether the involved functions can be realized correspondingly.However,the completion quality of system function verification is largely affected by human factors.If the tester’s test experience or analysis ability is insufficient to effectively analyze the faults generated in the process of system function test,it will have a great impact on test results.Therefore,how to realize the automation of system function verification and reduce the involvement of testers has become a research hotspot in urban rail transit industry in recent years.To realize the automation of system function verification,it is necessary to realize key technologies such as automatic generation of test scripts,automatic execution of test sequences and automatic analysis of test results.The acquisition and identification of drive machine interface（DMI）information is the basis of test result analysis.This paper focuses on the automatic recognition method of DMI interface information,including DMI images preprocessing,images feature extraction,images classification and identification.In addition,the application of the identified information in system function verification is discussed.The main work accomplished in this paper is as follows:1)Based on the analysis of the research status at home and abroad,the structures of CBTC system simulation platform and DMI interface are analyzed,the characteristics and categories of interface information are summarized,and the identification scheme of DMI information is proposed.2)The location method of DMI interface and the identification method of different types of information are studied.The positioning of the DMI display area is realized by dynamic contour detection technology,the recognition area is divided by the combination of Canny edge detection and Hough transform.Through the above method,the positioning difficulties caused by camera erection angle and image shaking can be solved.According to the characteristics of DMI icons information and characters information,the decision tree method is used to identify icons,the K-nearest neighbors algorithm is used to identify numbers and letters,and the Tesseract OCR is used to identify the Chinese characters.3)Based on Visual Studio development platform and python assembly language,combined with OpenCV computer vision library,the development of DMI information extraction and recognition system is completed,and the application of the recognition system in function verification is discussed.The identification result of DMI information is constructed as actual train status,and the relevant information in test cases is constructed as expected train status.According to the comparison results of the two train status,it is judged whether the corresponding function of CBTC system is realized.