Three-dimensiOnal Passport for process Individualization in Agriculture (UTOPIA)

Agricultural food products naturally vary in their detailed internal structure. To facilitate early detection of health risks due to contamination or diseases, predict maturity state and minimize waste, it is crucial to take the internal characteristics of each individual product into account as these allow assessment of product quality, individualized product processing and detection of harmful specimens.

The UTOPIA project will develop a powerful approach for in-line 3D X-ray scanning of agricultural products, which will enable the creation of an individual digital passport of each product containing features related to safety, quality and morphology derived from the product’s 3D internal representation. Our approach is based on advancing and combining two key technologies: (i) high-throughput, simultaneous 3D imaging of groups of products by multi-spectral X-ray tomography, and (ii) data-driven feature detection using deep learning techniques. By creating automated tools that link the morphological information to functional properties, decisions can be made in real-time.

Our approach will pave the way towards fully individualized quality assurance and processing of agricultural products, raising the bar for food safety and waste reduction through early detection of disorders and diseases, and increasing value through individualized processing operations.

The UTOPIA project team combines world-class expertise of CWI on 3D image reconstruction and machine learning with in-depth knowledge on agricultural products from Wageningen University and Utrecht University and vast experience in high-volume production systems for processing fruits (GREEFA) and poultry (Meyn). We will jointly develop concrete demonstrators for bringing these smart technologies into the real-world high-volume setting.

Contactpersoon
Prof. dr. K.J. Batenburg

Nieuwsbrief

Schrijf u in voor onze nieuwsbrief en blijf op de hoogte van het laatste nieuws over Commit2data