OCR-Image-based Identification APP For Bank Card With Optimization And Implementation

With the rapid development of the social economy and the increasing status of the electronics industry,the importance of smart phones in people's life has become prominent.Cashless payments are based on the development of mobile payments,which are inextricably linked to the use of smart phones.The bank card needs to be bound to the mobile payment software,while the reading of the bank card number replaces the manual input step.However,it becomes a technical problem how to quickly and accurately read the bank card number,which attracts scientists to research and discuss on it constantly.Based on OCR image processing technology,this thesis studies the bank card number reading technology.Firstly,it introduces the origin,development background and research significance of OCR technology.Then the content arrangement of this thesis is introduced.After that,studies and compares each module implementation technology and the improvement of technology for some modules.Finally,the open source version of Android,methods before optimization,and current popular commercial software features are compared to the optimization results of this study.A brief introduction to the modules mainly involved in this thesis is shown as follows:Process optimization is to optimize the processing of the image from the process according to the characteristics of the bank card,to reduce the information of the image processing,to save the time of the bank card image processing,and to improve the processing efficiency.The improved binarization method is used to optimize the image processing.The image features of the image are maintained as much as possible after the color pixels are removed.The features of the target area are highlighted.The accuracy is improved of the image processing.The filtering technique is used to process the image to reduce the influence of noise on analyzing the image features and acquiring the target.The noise exists in scattered and isolated points.Compared with the large area,these noise points contain few pixels.The filtering is used to eliminate noise and it improves efficiency and accuracy.Combined with the method of expansion and corrosion,the processing of worn bank card pictures is enhanced.Character cutting is used to divide the bank card number,then the bank card number is composed of a series of numbers.However,the digital grouping may have different forms,so the method of cutting the characters is not the same,and the vertical projection cutting method is realized.Therefore,the cutting accuracy is improved.In order to improve character matching,the matching algorithm is improved according to the characteristics of numeric characters.The accuracy and efficiency of character matching are also improved.The innovation of this thesis is that it improves the accuracy and efficiency of the program reading bank card number and bank information by improving the image recognition and reading method by improving the graying method,binarization method and character recognition matching method.