The Design Of Paper Currency Recognition System

RMB paper currency anti-counterfeiting and sorting play an important role in Chinese financial systems and commercial entities.But it is fairy weak in the domestic research in these aspects.Based on the theory of image processing and pattern recognition,algorithms about RMB anti-counterfeiting and sorting have been presented in this paper after research into RMG paper currency characteristics,which include algorithms of par value and orientation identification,serial number detection.Image location,tilt correction and image segmentation are included in image preprocessing.Due to a high real-time request for the paper currency anti-counterfeiting and sorting system,traditional edge detection algorithms are unsuitable for boundary location,So in this paper,the paper currency boundaries are determined by collecting finite points on the boundaries and then applying the linear least-square fitting method.On the other hand,the tilt correction is achieved by coordinate transform with the center of the paper currency as the origin.Dimension identification method is used to obtain the par value of RMB according to the fact that is a great difference between RMB paper currencies of different par values.And par orientation is identified based on the gray scale distributions of the paper currency.Binary image,de-noising,character segmentation,thinning,and recognition are included in the serial number detection.The binary image is realized by double setting threshold;the de-noising is based on the areas of connected regions;the thinning is achieved by morphological methods;and the character recognition is accomplished applying the contour features of the currency characters.The experiment show that the algorithms of this paper are of high accuracy and rapid speed.