Research And Implementation Of Banknote Scan Image Recognition Method For Digital Copier

The development of copiers has a history of about 70 years so far,the product has developed from a single function of black and white analog copier to today's multi-functional color digital copier.With the continuous progress of science and technology,the photocopying accuracy of the copier and the degree of image restoration are constantly improved,bringing people higher quality services,but also such problems like counterfeit money manufacturing.At present,the common foreign brand photocopiers on the market all have mature paper money identification technology,but due to the foreign technical blockade of China and the lack of our own research level,the domestic copiers in the field of banknotes identification technology research is not in-depth.Therefore,the research of the identification technology of domestic copier banknote should be carried out,so as to enhance the identification application of domestic copier banknote,enhance the security and controllability of domestic copier,and improve the market competitiveness of domestic copier.After carefully analyzing domestic copiers hardware architecture and software on the basis of workflow,under the copier control platform,the research of related technology of banknotes recognition are studied,the banknote scan image recognition method is designed and realized in the ARM control system,and the implementation results are integrated into the existing copy machine control platform,make us independent research and development of the copier with copying banknotes image recognition.This thesis firstly analyzes the hardware platform and control software of the self-developed copier,and proposes the design scheme of the scanning image recognition module of the copier according to the application requirements of the copying banknote recognition.Secondly,on the basis of the research and comparative analysis of spatial and frequency filtering algorithms,a new banknotes scanning image filtering method,which is suitable for the control environment of the copier,is proposed and implemented in ARM embedded environment.Thirdly,on the basis of studying the texture feature extraction algorithm of typical images and based on the two-dimensional Gabor transform,a band-pass filter bank with spatial locality and directional selectivity is designed to extract the texture feature of banknote images.At the same time,based on the extracted features,a banknote image recognition method with Euclidean distance as the classification parameter is designed.The effectiveness of the recognition method is verified by the ARM embedded programming.Finally,the banknote scanning image filtering module and recognition function module are integrated into the control system of the copier developed by the unit,and the copying test of the integrated system is carried out.The test results show that the thesis designed and implemented banknote image recognition method has high accuracy and operation efficiency,and meets the design requirements.In this thesis,the banknote image recognition technology can be applied to the control system of the domestic copier to identify and prevent the banknotes copying,which has certain significance to improve the core competitiveness of the domestic copier and the security and controllability of the copy.