Artificial intelligence for sustainable quality assurance of fresh food

The project is developing a multimodal AI assistance system for automated quality monitoring of fruit and vegetables in food retail.

The aim is to use edge AI (local AI analysis directly at the shelf) and federated learning (decentralized AI training to protect customer data) to dynamically integrate environmental factors and replace traditional laboratory tests. Machine learning methods, such as transformer architectures and multimodal foundation models, should ensure a high level of accuracy even with low-cost sensor technology. In addition, explainable and trustworthy AI methods are being developed to ensure a robust and comprehensible freshness assessment.

Intelligent image recognition systems are used to record optical data and assess the quality of the products. In addition, compact gas sensors are installed directly in the fruit crates to detect volatile organic compounds (VOCs). By combining the two approaches, the project aims to enable precise predictions of product quality, thereby contributing to the sustainable use of valuable resources and helping retailers to work more cost-efficiently.

With its research, the project aims to contribute to the reduction of food waste due to spoilage in the wholesale and retail sectors.