Research for the intelligent RMB sorter based on ANN

A smart and cost-effective intelligent RMB (Ren Min Bi) sorting system is built by theoretical approaches and technical facilities integrated of electrics, magnetics, optics, maths, mechanics. It can be used to identify authenticity, denomination, orientation, version of RMB, whether RMB is new or old, and count the number of sheets and the amount. Real money and false money can be distinguished by three detection facilities including fluorescence detection of forgery, infrared detection of forgery, and magnetic detection of forgery during every working process of the sorter. Test shows that the accuracy rate is 100%. In above process, once false coin is detected, the system immediately provides sound-light alarm, then false money is sent to the lower output (real money is sent to the higher output) simultaneously without interruption of the operation at that time, which is the feature of the sorter. Classified identification intelligence technology and picture identification processing technology based on ANN (artificial neural network) are skillfully combined with four relevant algorithms including BP (back propagation) -ANN, LVQ (learning vector quantization) -ANN, Kalman, genetics, then highly automation and intelligence of the sorting system are successfully realized, which is another feature of the sorter.