Research On The Image Acquisition And Pre-processing System Of Paper Currency Numbers Based On ARM

The serial numbers on the paper currency are the amount identifiers of the paper currency issuance, and each sheet has its own serial numbers and the same numbers can't be used twice, so the serial numbers on it can identify paper currency's identity. Developed an intelligent system by computer vision for paper currency numbers recognition, the numbers are recognized and recorded with binding by the system, then the organization of the paper currency could be easy. It is propitious to the detection of hijacked accidents and recognition of counterfeit paper currency.In this paper, a sort of self-designed image acquisition and pre-processing system of paper currency numbers with CIS(Contact Image Sensor)based on ARM and CPLD is introduced. It brings out a real-time approach for collecting high-speed image information and image pretreatment. The main contents of this paper are as follows:This system is mainly designed of hardware. The collected CIS image information is saved in the SDRAM(Synchronous Dynamic Random Access Memory)after some pretreatments such as occulting output compensation, secondary filter, analog-digital conversion and binarization. Then the image information will be further processed with ARM. The time driving signals of CIS and the handclasp signals of the system are both designed by CPLD.Additionally, lots of efforts are done on the software design of the system. For instance, the design of bootloading programmes about ARM is realized. Secondly, the image signals are transmitted to the PC through the UART port. Subsequently, recurrence of the image, orientation of the numbers'region of paper currency, and smoothing and filtering the binarization image about the paper currency are completed. In the meantime, the algorithms of character recognition are studied, and one kind of character segmentation and recognition algorithm is chosen which is suitable for our system,and the design thought is put forward as well.The new technical innovations of this essay are: using the hardware circuits to realize the binarization of image signals to reduce the time consumption of running the recognition algorithm; designing occulting output compensation circuits to improve the CIS image quality; choosing ARM as CPU to guarantee the real-time characteristic of the system; adopting the sea-content memory SDRAM to solve the bottleneck problems between the memory space and the running speed of programs.The experimental results reveal that the image of binaryzation shows very clear, and the aim of pretreatment is achieved. This research output can be extended to not only the entry system of securities and bill character, but also other field's character identify system. It has a wonderful prospect in the market and so comprehensive engineering application value.