A Currency and Denomination Detection System using Raspberry Pi and Machine Learning

The development of the note detection system is outlined for an embedment using a Rasp- berry Pi as the base for the work home with machine learning algorithms. The first domain of this work is currency identification that involves analyzing the amount depicted by a currency note and identifying the country to which the currency belongs to. This system is very applicable on automatic telling machine or better still ATM, vending machines, and currency exchange services since there is need to identify many types of currencies in the shortest time possible. Issues arising out of practice in this area include handling numerous and diverse currencies in relation to aspects such as size, color and security features among others. The last important problems are those of identifying between the real and fake coins and notes – the task to be accomplished by the system should demonstrate high accuracy. Furthermore, the recognition must be possible regardless of lighting conditions, and the system should be able to recognize partially damaged notes, as well. Some of these difficulties call for efficient image processing and machine learning strategies with a dependable high accuracy level.