Deep Learning, YOLO and RFID based Smart Billing Handcart

Abstract

Someone visited mall to buy the items they need on regular basis and pay for them. It is needed to verify how many goods are sold out and produce customer’s invoice. When someone visits a store to purchase goods, he/she must exert effort to choose the appropriate items. Furthermore, it is stressful to wait in queue to get the invoice after that. Therefore, we are suggesting a smart cart system created using RFID to keep track of the products that have been bought, as well as online billing transactions. The system will also make product recommendations based on user purchasing patterns from a centralised system. Each cart will be equipped with an RFID reader, and every item in the store will have an RFID tag as part of this system. Centralised systems will be used for web transactions and recommendations. Additionally, an RFID reader for anti-theft purposes will be located at the exit entrance. The proposed system includes a camera that uses deep learning to identify the object and a load cell that weighs the object that is fastened to the shopping cart. The system will generate the bill when the customer scans the item in front of the cart's fixed camera. Numerous methods can be used to implement object recognition. Methods like R-CNN and YOLO use area proposals to create bounding boxes, which are then used to run classifiers throughout. The duplicates are then removed using post-processing methodology. We utilized the YOLO paradigm because of this.