Research And Design Of Modular Coin Detection And Statistics System

Money of small denominations is indispensable in daily trading with the change of settlement, is a kind of the most frequently used currency. Compared with small denomination notes,coin resistance to wear、materials easy to recycle use again、 reduce the spread of the virus and so on.According to statistics, a small coin is 100 times the service life of the small denominations, and can recyclable use again. Therefore, Replace small banknotes with coins is adopted widely in the world, and is also the China’s central bank efforts to force into the goal.This no doubt will further expand the usage frequency of the coin.But, along with the COINS in circulation is more common and frequent, followed by such issues as the proliferation of counterfeit, counting the time-consuming to increasingly prominent.This paper studies these problems and more popular in today’s machine vision inspection, we proposed a modular design and the denomination of the coin detection system statistics. Compared with more common for a single coin recognition and detection research, this research on an image containing an image detecting more coins. So that we can achieve mass denomination coin recognition and statistics, to expand the use side of the system, it is expected to be freed from the tedious inventory work.The system includes an FPGA-based front-end and back-end image pre-processing module PC-based algorithm identification module. Front-end modules with Altera’s FPGA (EP4CE10E22C8N) as the core processing chip, high-speed parallel processing capabilities depend on FPGA hardware acceleration.Mainly to complete the function:configure the image sensor and the YUV422 format screen image acquisition, image gradation processing, median filtering, binarization. PC-based back-end module identification algorithm processing tasks are: segmentation, image pixel size and features of a single coin Hu extraction initial recognition based on the pixel area, the secondary image recognition based on BP neural network, COINS denominations statistics, etc.Choose USB Interface Communication between the two modules, USB interface chip is Cypress’s Cy7c68013A widely used, Has the characteristics of used widely(almost every intelligent platform),communication speed, plug and play,etc. Comply with the modular thought. So in accordance with the development of inter-module functionally independent, easy to modify independent modules, such as adding more processing functions on the front-end FPGA modules or algorithms to achieve more recognition at the upper end of the module, and in the future under the front-end FPGA module processing functions remain unchanged, The rear end PC module processing algorithms ported to other small core chip (ARM or DSP) embedded processing platform can be extended to the use of the environment on small mobile devices, such as public transport and vending machine.Finally, the whole system alignment test shows that on the basis of certain general pseudo ability, can complete the expected requirements for the coin detection and the total statistics from a image of a multi coin image.