Identification Algorithm For The Banknotes Magnetic Security Line

With a wide use of the5th generation of Chinese banknotes, various forms ofcounterfeit banknotes emerge in endlessly, and intelligent counter is becoming more andmore significant, where its discriminating ability directly impacts on the circulation of thebanknotes. After years of technological development and its’innovation, the identificationof banknotes extended from the initial fluorescence approach to magnetic method,especially the fifth generation RMB embedded with machine-readable magnetic securityline, promotes the continuous development of the magnetic discriminating technology.This article studies magnetic security line which is very stable signal, analyze its magneticimage features, and identify which denomination banknotes the feature of magneticsecurity line belong to.The indentification algorithm contains four approaches, theres are preprocessingoperation, magnetic security line locating, features’ extraction and classification.Pretreatment mainly is treated for image noise reduction and its magnetic gradient imagebinarization in order to obtain its initial orientation. This paper propose a method toquickly calculate the tilt angle of the security line where the initial orientation needs tiltcorrection to accurately extract the security line. This method can further identify theexistence of magnetic imaging magnetic security line. Finally according to the accurateextraction of the security line we can get its characteristic waveform based on verticalprojection and then get its feature and then SVM algorithm is used to classify this feature.The study makes clear that this feature can be well expressed in differentdenominations of magnetic security line, and precisely classified. In the experiment30%of small specimen were selected for training, and100images were taken to be classified,where the correct classification rate is99%.