The Study Of Currency Number Recognition Based On Characteristics

At present, all kinds of tickets are widely used in people's daily life, such as the bankmoney, checks, tax invoice and prize securities, out of necessity, the tickets are all printed.Paper money as a kind of important bills, their number is unique, without repetition,therefore it can be used as a unique identifier. Using image processing and patternrecognition theory to achieve bank note number identification records to some transferequipments, for example: ATM machines, cash registers, sorter and so on. Its an effectivemeans to prevent counterfeit money circulation and money laundering.Compared with other numbers, paper currency number is with small characters, longdigits, not fixed size, diversed characteristics, additionally, recognition system requireshigh real-time, accuracy, so using the traditional method of template matching recognitionis often required the character size normalization, consume more time, so the traditionalpaper currency number recognition is not suitable for practical application.This paper used a paper currency number recognition Based on characteristics, First,analysis the acquised image data, correct and zoom processing the image, due to ordinaryinterpolation scaling is with large amount of computation, will take a long time, so we useimproved immobilization with multiples of the bilinear interpolation method to enlargeimage, Experiments have proved that it can significantly improve the computationalefficiency. After processing the two values of images, do segment to each image character,using horizontal projection, vertical projection, then horizontal projection again, to achieveeach character positioning segmentation. Then do character classification recognition foreach character using the designed method. Character classification algorithm is based onthe use of character features of each character, do matching on the local features.Calculation amount is small, matching degree is high. Through the computer softwarerealized RMB and euro recognition, experimental statistics show that this method has goodrecognition effect and running speed, can meet the requirement of industrial production.