Smart Cash Register

Abstract

Currently at grocery stores like Publix, there is a lack of systems that employ machine learning to keep track of inventory. One particular area where inventory management can be problematic is at the checkout. The goal of the project is to create a “smart” cash register that can scan, weigh, identify and price various produce items through the use of deep learning classification. This “smart” cash register will make the jobs of cashiers easier, removing the need for memorizing various codes. It also simplifies the back-end record keeping that the store does. Each transaction and sale will be recorded to a database, further removing the need for self-reporting that some stores have implemented. This project will be accomplished using an NVIDIA jetson nano as the computer, a Hx711 load sensor scale and a webcam to view and scan the items. The camera will be positioned towards the scale in order to scan and weigh simultaneously.