Innovating Vending Machine Through Object Detection for Currency Recognition and Enhanced Cash Handling

Advancements in technology have transformed vending machines into active and autonomous devices. In recent years, vending machines have become essential as a standalone and low-maintenance system because malls and stores are not open 24 hours a day. However, in countries such as Pakistan, widespread adoption of vending machines is limited due to a lack of local manufacturing support. This study introduces a cutting-edge vending machine in order to establish an affordable, pioneering embedded system. This research presents a bill acceptor that employs a transfer learning model of object detection (FOMO) for currency recognition and authentication through an ESP 32-CAM within a mechanical cash transit system. The commercial bill acceptor, which is not readily available in Pakistan, can be replaced by the proposed bill acceptor. By offering a bill acceptor as an initial approach, which has a 71% accuracy rate through edge impulse testing and realtime trials, this study seeks to revolutionize the vending sector. This prototype is the first step towards overcoming a fundamental obstacle in local manufacturing of vending machines.