Research And Design Of The Paper Currency Classification System Based On Dual Core Processor

In recent years, with the rapid development of modern economy, worn, defaced money in the currency in circulation is becoming more and more. It hindered the circulation of commodities, and caused inconvenience to people’s daily life. But the work that Classification of the coins is still largely rely on manual processing, the disadvantage that Low efficiency, poor quality caused by manual processing brings a heavy burden to the financial system, Therefore, banks and other financial institutions need a high-performance Financial Equipments, especially the urgent demand for the Paper-Money-Classification-Equipment (PMCE), and the PMCE is showing its huge economic and social benefits.The paper carried out a detailed analysis and discussion about the Image feature of the paper money and the Image processing theory at first, and then Analyzes and compares the advantages and disadvantages of the Traditional methods that used in the paper money classification, and proposes a new algorithm about the paper money classification at last. The paper contains five parts as following:the classification of money’s face value, the classification of the worn money, the classification of money’s Orientation, the classification of money’s old condition, the classification of money’s edition.The algorism that classifying the money’s face value is based on the color recognition theory in the HSV color space, so as the classification of money’s edition, it determines the version of the money by Identifying the Color information of Crown Code. The classification of money’s Orientation is based on HMM theory, and the method is a statistical method. The algorithm proposes the idea of Block Histogram and it identifies the level of the money’s old condition.The algorithm that classifying the worn money is formed by joining the morphological filtering theory to the Background area statistical methods. And it used to remove the Small stains and reduce the calculation errors. Simulation results show that:The algorithms Mentioned above have the advantages that easy to implement, high fault tolerance, effective, and so on. And the algorism meets the requirements of the classification system.Two classification system platforms are designed and implemented in the paper at the same time, one is designed with Atom-Core, and the other is designed with the OMAP-L137-Core. Through the actual test and analysis, the system platform with OMAP-L137-Core is selected as the classification system Platform.It uses the latest dual-core processor chip OMAP L137as the core, FPGA as a coprocessor; the OMAP-L137processor is supported by TI’s Software system and with a high speed, Strong anti-jamming capability. So, the system designed in the paper has the more simple hardware circuit compared with the previous system, and it is low-cost, easy to extend, Easy to install and debugged, Practical tests prove the system is stable and fast.