Identification And Research On The State Of Railway Fasteners

Is composed of rail and sleeper rail,connection between the rail and the sleeper parts is fasteners,its main function is to make the rail firmly fixed orbit pillow,the state of the fasteners are in good condition,become an important factor of the safety of railway transportation.Good condition of fasteners should be intact,but due to long time,to experience stress environment evil slightly,the fastener missing may appear,become loose,produce the phenomenon such as crack,the phenomenon has become one of the reasons for railway safety threat.In recent years,China has been keen to develop intelligent railway fastener inspection system instead of artificial ones.Digital image processing is the most widely used method in the research of automatic inspection system.It has the advantages of high precision and fast inspection speed.Before the research on fasteners detection method,this paper use of domestic and foreign research present situation and methods of research,and for the positioning of the fasteners,feature extraction and state recognition algorithm on the basis of the comparison,finally determined the suitable for the algorithm of this paper.First of all,this paper adopts the method of "cross crossing" to locate fasteners,which is the easiest and has good stability and practicability.Then,the feature extraction of the fastener image is carried out,and the direction field algorithm that has been successfully applied in many fields is adopted,which can greatly reduce the calculation amount of the recognition of the fastener.Finally using the template matching method to be the direction of the test image and standard image in the direction of the field value matching,support vector machine(SVM)is used to set the template threshold,the state of fastener,and the pattern recognition,this algorithm has high recognition accuracy,computing speed and simple calculation process and other advantages.This subject adopts LabVIEW PC programming,through the module structures,the above algorithm experiment,the validity of the algorithm are verified,implements the fastener localization and recognition work,good results have been achieved.The experiments show that the proposed algorithm has good stability and robustness to the recognition of fasteners.