Towards optimization of tomato cultivation using a digital twin

With current advancements in sensor technology, growers need new information systems to process and use the wealth of sensor data. This will promote a data-driven, resource efficient, and productive tomato cultivation, while maintaining crop health. Digital Twin technology provides means to assist the grower with decision-making, however its use in horticulture is still in its infancy. A digital twin is a virtual representation of a real object – here a greenhouse crop production system – which is being updated by real-time information provided by sensors. In this study, we present a Digital Twin for tomato cultivation to supply the grower with information on climate, plant status, and the consequences for growth and resource use. The twin consists of a greenhouse climate model and a crop growth model, as well as sensors that continuously monitor climate, light capture, plant weight, transpiration, water and assimilate content, and organ dimensions. In the first two months of a tomato trial, the crop growth model was calibrated on sensor data and additional manual plant registrations. The model performed well in the subsequent months, and its plant state was updated using the non-invasive plant sensor readings in a feedback loop. In addition, we used the data to calibrate an existing 3D plant model with a multispectral ray tracer and photosynthesis module, to estimate the requirements for assimilation light and optimal light spectrum. In the future work, a DSS-digital twin platform will be developed to provide information to the grower through a dashboard, showing the status of the crop and future predictions. It will generate advice via a scenario generator, modeling the effects of specific measures and greenhouse-crop settings.