The Research Of Infrared Characteristics Identification On Banknotes Based On HOG

With the growth of China’s economy and the globalization of trade, the circulation of banknotes is increasing day by day. However, the currency sorter in China has just started, identification and authentication apparatus less trustworthy, and counterfeiters are rampant, the study and improvement of national recognition and discrimination technology.In view of the above questions, this paper by means of wavelet analysis, Gauss filter, gradient direction histogram method HOG algorithm in-depth study on the recognition and discrimination of infrared technology. I analysis of the infrared image features of banknotes; and chose reasonable methods of paper currency image processing. In this paper, the orientation gradient histogram method is applied to the feature extraction of paper currency, and the characteristics of the paper can be realized by using the mixed Gauss model and wavelet algorithm, you can note features to achieve better authentication. The recognition of the nominal value of the above mentioned above is carried out and the characteristics of the infrared detection of the notes are false. The experiments show that the recognition of the infrared and the effect of the detection is better.This paper reviews the research status of RMB recognition and authentication technology, introduces the research background, significance, the main work of this paper and the arrangement of the chapters. Secondly, according to the actual situation,the image pretreatment, including to background processing, Gaussian mixture model filter and wavelet processing. The notes were wavelet decomposition processing and filtering through the mixed Gauss model, through mixed Gauss model independent noise points and the processing of continuous noise points in the image, for the HOG template and feature identification notes banalization templates obtained.Finally, in the process of detection, the paper adopts the method of template matching to reduce the repeated calculation, and realize the fast and effective detection. By experimental verification, the banknote recognition and identification technology, with advanced can get very good results, to meet market of sorter and sorting technology requirements and has a higher practical value.