Research on paper currency recognition by neural networks

Paper currency recognition is widely applied in many fields such as bank system and automatic selling-goods system. How to extract highly qualified monetary characteristic vectors from currency image is an important problem needing to be solved now. It is a key process to select original characteristic information from currency image with noises and uneven gray. This article, aiming at the specialties of RENMINGBI(RMB) currency image, puts forward a method using linear transform of image gray to diminish the influence of the background image noises in order to give prominence to edge information of the image. Then the edge characteristic information image is obtained by edge detection using simple statistics. Finally by dividing the edge characteristic information image in the width direction into different areas, getting the number of the edge characteristic pixels of different areas as input vectors to neural networks (NN), carrying out sorting recognition by three layer BP NN, paper currency is recognized. The proposed method of extracting input vectors has advantages of simplicity, high calculating speed, obvious characters of original image, good robust to different kinds of RMB. By experimental tests, recognition ratio to the new printing style of 100 RMB, the old printing style of 50 RMB, the new printing style of 50 RMB, 20 RMB, the new printing style of 10 RMB is 95%, 99%, 99%, 92% and 98% respectively. And the results are satisfying.