AI technologies have long been an integral part of day-to-day working life at the REWE Group. They help to streamline administrative processes, support shop staff with routine tasks and ensure an even better shopping experience for our customers. One such technology is the VOGEL image recognition model – a visual fruit and vegetable recognition system.
This AI system, VOGEL, recognises fruit and vegetables at the checkout without the need for barcodes. The principle behind it is this: as soon as customers or staff place a product on the scales, the fruit or vegetable is photographed. Software then analyses this image using probability-based sorting and suggests up to eight possible products on the checkout display. VOGEL not only distinguishes between different varieties, but also recognises whether the produce is organic or conventional.
Thanks to the AI-based system, both customers at self-service checkouts (SCOs) and checkout staff can quickly and easily select the correct product with a single click. This process not only saves time and makes shopping less complicated, but also helps to overcome language barriers – particularly at SCO checkouts.
Malte Eckardt, Technical Lead at REWE digital
Malte Eckardt is Technical Lead at REWE digital and the initiator and developer of VOGEL. From the outset, his focus was on solving a specific problem in the market using modern AI. “The fact that this has resulted in a solution now in use at checkouts shows the potential that lies in simple ideas when you think them through consistently and persevere in the face of challenges.”
To ensure the system works reliably in practice, VOGEL relies on modern image recognition logic: the system uses what is known as computer vision and an underlying neural network (KNN). This is trained using images of a wide variety of fruit and vegetables, so that over time the system learns the typical characteristics of the products and can independently assign new images to the appropriate category.
Based on these continuous learning processes, the technology has been in use in PENNY stores since mid-2021, where it supports both manned and self-service checkouts (SCO). VOGEL now recognises 75 different fruit and vegetable categories here, comprising 340 individual items. At REWE, the phased roll-out at SCO checkouts began in January 2026. Here, VOGEL currently recognises 87 different fruit and vegetable categories comprising 574 individual items.
VOGEL is a system developed in-house by the REWE Group, originating from a hackathon organised by REWE digital. The resulting system once again underlines the Group’s pioneering role – particularly when compared with its competitors. REWE was the first company in the sector to introduce an AI-based system trained on over 100,000 images. This extensive database ensures an exceptionally high success rate in the recognition and classification of products. Furthermore, the system is designed to handle a wide variety of situations encountered in everyday shop and payment scenarios. This versatile innovation helps to reduce stock-taking and operational errors, thereby significantly improving process efficiency and accuracy in day-to-day operations.
