Application Of Digital Image Recognition In Enterprise Railway

With the progress of society and the rapid development of various technologies,image recognition technology has become increasingly widely used in all walks of life,which profoundly affects the operating methods and efficiency of enterprises.Digital image recognition also has its wide application fields and application scenarios in enterprise railways.This paper makes a preliminary analysis and Research on the recognition of railway freight car number and the identification of paper contents such as driver's declaration.The main work of this paper is as follows:? Describe the theoretical research and current situation of image recognition technology;? Discuss the key points of identification of railway freight car number:(1)In the aspect of string location of vehicle number,the method based on string shape with threshold is selected.(2)In the digital segmentation of vehicle number image,the projection method is combined with scanning vertical and horizontal pixel points.The projection method mainly determines the approximate position of self-characters,and the accurate segmentation is mainly realized by scanning pixel points.This improved projection method greatly improves the reliability and accuracy of character segmentation.(3)In the selection of recognition methods,the template matching method is used in pattern recognition,which improves the efficiency of pattern recognition.In the practical operation,this paper shows the interface of the car number recognition system and the code of the program in detail,and shows the results of the whole operation process of the system.(3)further expand the recognition object,use BP neural network to recognize the handwritten figure image,and do some research on the handwritten figure recognition,so as to lay a foundation for the recognition of paper content such as driver's report in the future.? This paper also makes a detailed research on handwritten numeral recognition,expounds its theory in detail,and realizes the handwritten digit recognition through program,which lays a solid foundation for the recognition of paper documents such as train formation sequence and driver's declaration.There are 37 images,2 sheets and 57 references in this paper.