Coin Identification and Conversion System using Image Processing

Abstract

Different features and template designs of currencies give travellers a hard time identifying and recognizing a coin; that is why a device to locate and recognize a coin is a must. Four currencies were used in the study. The researchers applied the YOLOv3 and CNN to create the system. Identification and recognition of coins is challenging with their many orientations and widely changing patterns. To resolve the problem, the researchers obtained the optimal distance of the camera to the coin and tested different angles of rotation. It was found out that a 6-cm distance from the camera to the coin is best for easy coin identification. The system’s accuracy are 98.15%, 98.15%, 97.22%, and 96.30% for 0-, 90-, 180-, and 270-degree angles, respectively.