Research On Image Characteristic Recognition Of Paper Currency

Paper currency recognition is a popular issue in the pattern recognition area in recent years. It is widely used in many areas. The paper currency identification instrument developed from this technology plays an important role in the finical system such as bank and so on.Image characteristic recognition is the essential technology of paper currency identification instrument. It mainly accomplish denomination identifying,facing recognition,old and new recognition,defect detection of the paper currency. The algorithm which used in the paper currency identification instrument should be with the character of strong reliability and real-time.In the software design,this dissertation takes the fourth(100yuan,50yuan,10yuan,5yuan) and the fifth(100yuan,50yuan,20yuan,10yuan,5yuan) paper currency image as the object which based on the gray level feature to achieve paper currency's real-time recognition using digital image processing technique,pattern recognition and neural network technology. These methods in this dissertation have a strong adaptability. After the steps of orientation, lean adjustment, measure of size and extraction of the image feature, we can recognize its value, version, side, direction and deformity. On the defect detection, this dissertation proposes a new method. The black regions in the binaryzation paper currency image which not incomplete are filtered by the method of mathematical morphology. The method is simple and practical. On the old and new recognition, a new method of fifth edition paper currency is proposed which based on the region of old and new characteristic region.This method is easily applied into practice.It is proved that this system can meet the technology requirements in real time and separation.