Kan kunstmatige intelligentie de productie voorspellen?


Can Artificial Intelligence predict production?

Published on
October 22, 2019

Talking about Artificial Intelligence there is no better display of this technology than what smartphones, social media, self-driving cars or video games are doing with it. What about demonstrating the capabilities of AI in greenhouses? Is it possible to predict cucumber harvest for the coming weeks? What information does AI need to make correct predictions? The Business Unit Greenhouse Horticulture at Wageningen University & Research is working on the development of an AI yield prediction model and associated database.

A greenhouse is a complex system with several components such as the crop, climate and irrigation strategies. Within this system sensors measure various plant characteristics with optical and imaging techniques. As a result the plant itself acts as a sensor of its own biological status and its environment. Nowadays, growers monitor the crop and decide on alterations of their greenhouse management to achieve production goals.

Capturing the intuition and 'green fingers' of experienced growers with sensors that continuously collect data can offer great opportunities. A database can be filled with meaningful and adequate data describing the status of climate and crop. Useful information in the datasets can be distilled and used towards data-driven decisions made with AI.

The GrowDat project aims to develop an AI framework that identifies the important climate and crop parameters for making accurate yield predictions. AI can support growers' decisions, better understand underlying processes and discover new patterns of the greenhouse production system.