Identification method of crown word number (serial number) of bank note
A recognition method and a serial number technology, which are applied in character and pattern recognition, computer parts, computing, etc., can solve the problems of ignoring the essential characteristics of serial numbers, the recognition rate cannot be reached, and banknotes cannot be recognized, so as to facilitate subsequent processing. , improve the recognition rate, help to monitor and manage the effect
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Publication Date
- 2014-03-12
Smart Images

Figure 1 
Figure 2 
Figure 3
Abstract
Description
technical field
[0001] The invention relates to a novel banknote serial number recognition method, which belongs to the field of banknote detectors. Background technique
[0002] Today, in China, where the economy is developing rapidly, the large circulation of RMB and the variety of currencies make the monitoring and management of RMB face increasing pressure, and the key to monitoring and management of RMB lies in the management of serial numbers. Therefore, the core issue is how to accurately identify the serial number. However, the banknote detectors currently on the market have a low recognition rate for serial numbers. Therefore, it is particularly important to develop an efficient and mature banknote serial number recognition method, which is directly related to financial security and order.
[0003] The traditional identification method of the serial number mainly focuses on a single method, ignoring some essential characteristics of the serial number itself, and t...
Examples
Embodiment
[0131] The following takes the fifth set of RMB 100 banknotes as an example to illustrate the specific identification process of banknotes. Firstly, the image of the face of the RMB containing the serial number is collected under visible light, and the area of the serial number is located. Using the serial number area as the source image, it is divided into two independent steps.
[0132] First: first use formula (1) to carry out wavelet transformation on the image, find out the modulus of wavelet coefficients, refine the image based on the local modulus minimum value, and finally obtain the skeleton of the serial number image. Afterwards, various features of the skeleton are extracted to describe the serial number. If the serial number is normalized to 24*18. Then take the serial number "9" as an example to extract its features: the number of bifurcation points is 1, the number of circles is 1, the number of sides is 1, the cumulative sum of the vertical crossing volumes ...