Research Of RMB Optical Character Recognition Algorithm And System Implementation

With the development of Chinese economy, the circulation of the Chinese bill is grow-ing as well. Chinese anti-counterfeiting field is facing a huge challenge when various foreignbills entered Chinese market.Therefore,the research on RMB optical character recognitionalgorithm and system implementation not only help the management and inspection of RMBcirculation,but also promote anti-counterfeiting techniques of China.The RMB optical character recognition system consists of two core techniques:an effi-cient RMB optical character recognition algorithm and a high-performance embedded hard-ware platform. This article proposed a RMB optical character recognition algorithm basedon image skeleton features, which has conducted extensive research in RMB optical charac-ter recognition algorithms.Then,this article designed a set of hardware platform and carriedout extensive tests.This article proposed a new method of RMB optical character skeleton feature ex-traction in the aspect of RMB optical character recognition algorithm and software opti-mization,which can improve accuracy of RMB optical character and performance of theplatform. Secondly, it develops a new method of DSP software optimization method basedon the feedback of CCS message.Finally,the overall performance of this system has beenimproved greatly by the code optimization.In the aspect of system implementation of RMB optical character recognition algo-rithm, this article analyzed the designment of software and hardware from the image pro-cessing engineering view, and designed each functional modules for the hardware plat-form based on DSP. The hardware system mainly consists of image processing, logic con-trol and image collection and the software module mainly contains DSP framework mod-ule,collection control module and so on.The experiment results have shown that RMB optical character recognition algorithmand hardware platform fully meets the requirements on the the accuracy and speed for aRMB optical character recognition system. Therefore, it is feasible and valuable.