Design And Research Of Currency Sorter System On ARM9 Platform

Currency sorter system is a complex system, It combines a variety of techniques. Its main function is to recognize face direction of the bills and banknote denominations, and identify the new currency and the old currency, The latest systems also contain related function of paper currency discrimination. Thus it can meet the selection for the actual currency notes in circulation and it can be able to replace the tedious manual labor requirements. With the recent decades, the rapid development of hardware technology and pattern recognition and relevant Discriminating technology, Using the comprehensive technology of sorting system began to appear in the ninety’s, and the recognition rate is gradually improving. Due to the constraints performance of related chip, The earliest banknote sorters have the disadvantage of low processing speed and the disadvantage of large and complex mechanical structure. In recent years, With appearance of the high performance of the high-end processing chip ARM series and the high performance image processor series,it makes use of relevant software and hardware technology to produce good performance and relatively cheap price banknote sorter systems increasingly becoming a possibility.This paper first reviews the current situation the development of Currency sorter system at home and abroad in recent years, To construct a set of the main control system circuit(including the S3C2416 X series chip and its peripheral equipment, Discrimination processing circuit and motor speed PWM drive circuit) system, The connection of system circuit through a common interface and HPI high speed interface of image acquisition and processing system of chip. Subsystem contains its general peripherals and through the chip-specific video port(Vidio Port) and two contact image sensor(CIS, Contact Image Sensor), A/D converter, and conditioning circuitry of the image acquisition and processing platform.32-bit processing ARM9 S3C2416 X chip can ensure the timely summary and analysis of the banknotes data, High speed TMS320DM642 DSP chip of Tl company ensure real-time data acquisition and processing, and it can be able to complete a variety of algorithms of task requirements. Using CPLD(Complex Programmable Logic Dveecis Complex Programmable Logic device) finish the mission of Logic control design. The hardware setup allows the sorter system to overcome the hard real-time requirements, Thus it can complete the functional requirements. One of the hardware of the system innovation is the main control system upgrade from the traditional single-chip microcomputer to the core of ARM system, its interface expansion and convenience of strong performance, also has a larger upgrade. Second, the subsystem uses a high-speed dedicated image video DSP chip,and it has a dedicated image processing interface. The third is that the system uses a dual CIS sensor design.Then we do some preliminary simulation study to the the algorithm for the system. Among the algorithm for the system,the paper currency discrimination make use of paper money special region texture analysis based on wavelet packet. The basis of the method is that the real parts of the texture of the note has its unique properties. And the comparison of the wavelet packet transform and wavelet transform, the wavelet packet transform does not exist the shortcomings of the higher spatial resolution, the lower frequency resolution. Thus we do the simulation study,and the study which we do can achieve a certain effect,it also has a certain practical application prospects.At the end, Neural network patterns are used in the recognition of the face of bnaknotes and the values of bnaknote and other information. Two neural networks were introduced BP(Back Propagation) neural network and LVQ(Learn Vector QuantiZatoin) neural network, And we simulation based on two kinds of neural networks. Through the analysis, we select the LVQ algorithm to identify the 05 edition RMB. By adding Gaussian noise to simulate the old banknotes and new banknotes, a certain degree of simulation is done. In the end of this paper, it puts forward the ideas and suggestions for improvement.