AI, machine learning and synthetic data infrastructure for simulation

AI is transforming the way agriculture and horticulture are studied and applied. Using digital twins – virtual replicas of greenhouses, fields or crops – we can test and optimise cultivation strategies without real-world risks. We also generate and apply synthetic data to train and validate models, even when real-world data are scarce or costly to obtain. Examples include virtual breeding populations and image recognition for quality control. That way, we can accelerate innovation.

Our work is highly multidisciplinary: plant scientists, data specialists and AI experts collaborate to create solutions that directly benefit practice. Responsible AI is central to our approach — we ensure data quality, transparency and ethical integrity.

• Data acquisition and scoring: Wageningen University & Research provides the infrastructure to generate the right data through an integrated sensor approach — combining sensors, lighting and camera positioning — and to design the most effective experimental setups.
• Digital twins in practice: By creating digital replicas of greenhouse and arable systems, WUR can simulate multiple scenarios. This enables growers and businesses to optimise harvest timing, irrigation strategies and climate settings.
• Synthetic data: WUR generates virtual datasets that train and test models without relying on extensive field trials.
• Decision support tools: Our AI models turn complex data into practical insights, for example on water use, yield forecasts or long-term crop management strategies.
• Responsible AI: We design and work with transparent and explainable models in close collaboration with stakeholders, preventing bias and ensuring user trust in AI-driven outcomes.