Implementación de un prototipo de máquina clasificadora y contadora de monedas basada en visión artificial.

Abstract

It was carried out the design and implementation of a prototype coin sorting and counting machine that uses artificial vision as a method of identification and counting, as well as a mechanical system for classification, establishing a new technique for money counting operations. Commercial machines with high speeds and multiple functions integrate patented sensors for the identification and counting of coins, so the camera in the process of artificial vision is a low cost element that enmulates these sensors. The processing of the images is done with the Asus Tinker Board S card after capture with the Raspberry v 1.3 camera and then transmits the results obtained to an Arduino Uno through the 12C communication protocol. This second card is responsible for managing the processes of the motor speed control in the system for the separation of coins and the HMI interface. It has a simple interface with three buttons with which someone can browse through the menu and view the data obtained during and after each operation with the coins. The tests carried out established a result of 0% error in low speeds, defining artificial vision as a slow technique in the designed prototype. The cost of implementing the prototype remains relatively low with respect to the acquisition of a commercial machine, as it does not have a national production of this type of tool. It is recommended to use material with greater rigidity in the mechanism of separation of coins to avoid jams and extend the useful life of the prototype.