Smart Shopping Trolley System

Abstract

Many efforts have recently been made to reduce the amount of time it takes to pay in different retail settings. Furthermore, as the Fourth Industrial Revolution progresses, artificial intelligence research advances, and IoT computers become more portable and less expensive. As a result of combining these two innovations, it became possible to create an unmanned environment on behalf of humans in order to save users' time. We present a smart shopping cart device built on low-cost IoT hardware and deep learning object recognition technologies. Initially a Raspberry Pi and an Rpi Camera are used to detect the products and afterwards a GUI is offered for displaying and calculating the total bill of the products being added inside the smart trolley, and a deep learning server where learned product data are processed make up the proposed smart cart framework. The customer will review the list of products placed in the smart cart through the GUI. The smart cart device proposed can be used to create low-cost, high-performance unmanned stores. Moreover, the number of classes that we have in our dataset are 17. Our mean average precision (mAP@0.50) is 98.60%. Our Precision value is 0.98, recall value is 0.99 and our F1-score = 0.98.