Machine for counterfeit banknote detection

This project aims at developing a fully working machine for classifying, validating and analyzing banknotes in order to detect fake, counterfeit and defective banknotes at a high working speed.

This project has the final aim of developing innovative classification algorithms to detect counterfeit banknotes and classify them depending on the defects. Correct  banknotes must then be recognized (with the correct currency and type of banknote). The developed algorithms must account for the acquisition through a CIS (Contact Image Sensor) and the consequent skew of the acquired note. Moreover, the side of insertion must be recognized. As an even more challenging task, the system must work at very high speed (a hundred of banknote per minute) and on an embedded device (a standard no-floating-point DSP).