The Research On Banknotes Recognition And Discriminating Technology Of Infrared

With the growth of our economy and the globalization of trade, exchange ofnotes between countries has become the essential business of Chinese banks. However,sorter is just started in China, identification and authentication apparatus lesstrustworthy. Plus the counterfeiters are rampant, to study and improve the papermoney identification and authentication technology.According to the actual situation, first of all to the image preprocessing,including CIS non uniformity correction, the correction of angle and image filtering.Compare the theory of CIS non-uniformity correction algorithm; according to thereal-time, select suitable method. The correction of angle, based on the Houghtransform and a methods based on spatial transform are compared, according to theactual situation to choose the tilt angle correction algorithm. Similarly to the linearfilter and nonlinear filter are compared.In feature extraction process, compared with the Principal Component Cnalysisand Non-negative Matrix Factorization of. The use of the difficulties andcharacteristics of the two methods were experimental. The principal componentanalysis, the number of principal components selected needs a lot of experiments. Fornon negative matrix factorization, decomposition of two matrix size requires a largenumber of experiments. In the classifier design, through reading a large number ofdata, understanding to multinational paper money is very difficult by the singlealgorithm quickly. The paper packet recognition, and achieved good results.In this paper, the research status, the rupee, the euro currency at home and abroad,analyzes the visible light image features of these notes, difference and the infraredimage, use the reasonable method for image processing. In this paper, non negativematrix is applied to the notes feature extraction, through the Gauss model of therecognition of paper currency. Recognition results are good. In recognition of thecurrency notes known image, value oriented and conditions, notes infraredcharacteristic identification. Through the experimental identification andauthentication is better.In the authentication process, for identification of known notes, can reducerepeated calculation, which can achieve fast and effective identification. To identifythe authenticity, method is used by comparing the correlation coefficient and distancemeasure.The experimental results show that the banknotes identification and authentication technique adopted in this paper has certain advancement, can get betterresults, is able to meet the market requirements for the sorter sorting technology, hashigher application value.