The Design And Implementation Of Detection System For Railway Track Belt On Machine Vision

In recent years,transportation industry of China develop continuously,technological innovation and economic need promote the rapid development of high-speed railway,railway track laying quantity and mileage is increasing so that the burden of the daily inspection work of the railway track geometric is hard,the current railway track detection still remains in the manual stage,so this detection has strong instability and high failure rate.On the other hand,there is a big risk of railway safety inspection personnel.With the development of technology and the improvement of automation degree,the railway track detection will be developed to the trend of automation and intelligence.With the rapid development of machine vision technology,it has been applied in many fields,such as medical,manufacturing,security,military and so on.This paper presents a railway track detection system based on machine vision,focuses on the detection of railway rail gap distance and the lack of fastener,high speed of industrial camera and high precision of image processing technology replace manual inspection,this detection also has advantage of high reliability,data retention and wide applicability.According to the characteristics of the Qin Huangdao railway,this paper designs a new railway detection algorithm.Rail gap in captured image is detected by histogram statistical information after preprocessing.The suspectable image is processed morphologically,then calibration is applied to convert image coordinate to rail gap distance in real world.As for the fastener detection,the standard rail fastener image is processing by edge extracting and morphologically,finally a pattern recognition method of pixel scanning is designed to judge existence of rail fastener.This paper introduces the hardware system of the detection system,and the acquisition system can transmit the data to the processing system.At the same time,it also introduces the basic functions of the compiler and the software interface.After a large number of field data processing,the accuracy of the detection algorithm and the system is verified,and it can be used for reference in the same field.