Validity currency detector with optical sensor using backpropagation

Abstract

This paper refers to the problems that arise in everyday life, where most people still have problems in distinguishing between real money or fake money, especially blind persons. This paper will be studied how to make a portable tool that can read the signs visible on the money and then provide feedback in the form of sound. Ultraviolet rays will be used to emit visible image that is on the currency, then the image will be captured by the webcam as an optical sensor. A Single-Board Computer based Raspberry Pi is used to process the data that is captured by the sensor, then Single-Board Computer used to process data sent by the sensor, then the data is converted into histograms and compared with a sample using a histogram comparison, after comparison of data obtained later Single-Board Computer provides signals such as voice if the money is genuine or counterfeit, along with the amount. The tool is able to read and authenticity nominal money effectively, the results of the experiment resulting error value is less than 5%.