A sorting machine-based recognition method for multi-national banknote serial numbers

Abstract

The embodiment of the invention discloses a method for identifying serial numbers of multi-country banknotes based on a sorting machine. The method identifies characters to be identified through character matching templates. First, after converting the image of the banknote serial number into a grayscale image, the recognition area of ​​the character is determined through initial positioning, binarization, and re-positioning, and then the area size of the character is consistent with the size of the matching template through normalization, and then through Click and multiply to filter out matching characters and complete the recognition. This method not only avoids the influence of noise, but also can omit the filtering step; it is applicable to a variety of current mainstream currencies including RMB, US dollar, Euro, Hong Kong dollar, Japanese yen, etc.

Description

technical field [0001] The invention belongs to the technical field of automatic identification, and in particular relates to a multi-country banknote serial number identification method based on a sorting machine, a secondary segmentation method combined with prior knowledge of banknotes, and a character identification method based on template matching. Background technique [0002] In the financial field, each banknote has a unique serial number, which is equivalent to the ID card of the banknote, so the serial number becomes an important identification of the banknote. Whether it is classification or anti-counterfeiting, the serial number plays a vital role. At present, the number of economic disputes and illegal capital flow cases is increasing day by day. In addition, in the fight against violent banknote robbery and money laundering, the serial number is a powerful tool for the state machine to manage banknote flow, which is conducive to social stability. At present, ...

AI Technical Summary

Problems solved by technology

[0004] In order to solve the above problems, the object of the present invention is to provide a multi-country banknote serial number recognition method based on a sorting machine, which adopts a secondary segmentation method combined with prior knowledge of banknotes and a character recognition method based on template matching to solve the existing problems. There are banknote serial number recognition problems that cannot adapt to complex backgrounds and character size changes, and the recognition speed has increased by more than 8 times

Images