Exploration of an Indonesian Currency Legality Detection System by Utilizing Image Intensity of RGB Mean Values

Money is one of the objects used by the public to carry out legal buying and selling transactions in a country. The problem is there are fake bills that are printed by irresponsible people, so everyone needs to be able to know that the banknotes received is fake or genuine. But not everyone can detect the authenticity of a banknote, so a tool is needed to detect the banknote is genuine or fake. In this paper, software has been designed to detect the authenticity of Indonesian currency. In this paper, feature extraction of the grey level co-occurrence matrix with the features of entropy, contrast, correlation, energy and homogeneity is used to detect the nominal value of Indonesian banknotes and to detect the validity of Indonesian banknotes; the extraction of red-green-blue features with features is used mean R, mean G and mean B. The detected Indonesian currency was Indonesian currency from 2004-2016, with nominal values of Rp. 1000, Rp. 2000, Rp.5000, Rp. 10000, Rp. 20000, Rp.50000, and Rp.100000. The classification process uses Naïve Bayes.
From the test results, the system works well for reading the nominal value of Indonesian banknotes and detection the validity of the money can function properly.