Artificial intelligence and digital twins in postharvest food systems: transforming quality prediction, shelf-life management, and food loss reduction

Abstract

Postharvest food loss and waste account for 13 to 14 percent of global food production and nearly 400 billion US dollars annually, driven by the reactive nature of conventional quality monitoring, subjective manual grading, and fixed shelf-life labelling that fail to capture the dynamic environmental conditions experienced across modern supply chains. This review consolidates recent advances in artificial intelligence (AI) and digital twin (DT) technologies for postharvest food systems, an area where prior literature has largely treated intelligent quality assessment and digital infrastructure as separate research streams. It examines machine learning, deep learning, computer vision, hyperspectral imaging, and electronic nose and tongue systems for non-destructive quality prediction; digital twin architectures built on Internet of Things (IoT) sensing, cloud and edge computing, and blockchain traceability; and hybrid AI-DT approaches for dynamic shelf-life estimation and cold-chain optimisation. Applications across fruits, vegetables, dairy, meat, and seafood are reviewed, alongside industrial adoption barriers and an Industry 5.0 perspective. Integrated AI-DT frameworks enable continuous, product-specific shelf-life prediction that outperforms static labelling, supporting improved inventory decisions, reduced losses, optimised resource use, and lower carbon footprints. Realising this potential at scale requires standardised datasets, explainable AI models, interoperable digital infrastructures, and commodity-specific validation beyond controlled laboratory conditions. Falling sensor and cloud-computing costs are expected to extend adoption from large enterprises to small and medium producers, with future progress depending on integration with smart packaging, blockchain-enabled traceability, and autonomous decision-support systems, positioning AI-DT technologies as a central pillar of sustainable, climate-resilient postharvest management.