Serial Number Identification Of RMB Image And Application On Counterfeit Money Associated System

RMB is the legal currency of our country, just like our ID number, it's unique. Each of the RMB has only a specific serial number, it can be used as the identity of the RMB. Firstly, we study image recognition technology of RMB serial number systematically, and then focus on system on counterfeit money association. Finally, we apply this technology to the system.In study, we take the fifth edition of 100 yuan RMB as our research object mainly. Using the knowledge in the field of digital image processing and pattern recognition, we design the process of image pre-processing, complete location and character segmentation about serial number. Finally, we realize the serial number character recognition. In the image processing of binarization, we propose a new method. In this method, we superpose gray scale image before binarization. Experimental results show that character information is more prominent after using this binarization method, which lays a good foundation for the subsequent processing of characters. In addition, this paper also proposes an improved method based on a priori knowledge of rough location. Using this method, we can locate serial number area without some redundant information or missing part of character information. At the same time, we can carry on the character segmentation directly without the need for precise positioning. All of the above methods are all suitable for the RMB published in 2015.System on counterfeit money association is mainly using image processing technology, then analysis and calculate the characteristics of counterfeit money image, in order to judge the different counterfeit money is derived from the same or different crime dens. Through the analysis of counterfeit currency image features, we focus on the method based on single feature of counterfeit money association and then do test and analysis to it. Experiments show that the method based on the characteristic of serial number can achieve classification of counterfeit currency rapidly, and accuracy is higher than the other three features. Taking into account of the instability for the single feature of association, this paper will use the four single features to association according to a certain weight eventually. After experiments, it is found that the accuracy of the association is higher, and the stability of the result is higher than that of single feature association.Facts have proved that applies the serial number recognition technology to the system on counterfeit money association can realize the counterfeit money's association, and can make the source of counterfeit dens provide tangible evidence. Therefore, the technology applied into the system is a successful attempt. It can not only aid identification of counterfeit expert to recognize the source of counterfeit money, but also improve the work efficiency of the expert greatly. It also has a very important significance for the crime of counterfeit money.