Increasing automation in greenhouses could help tackle labour shortages and improve crop production efficiency. Marianne Curtis reports on how a Dutch university is challenging teams from around the world in this endeavour.
An international challenge trial which involves using artificial intelligence (AI) and autonomous greenhouses to improve vegetable production is nearing its conclusion.
The challenge was organised by Wageningen University and Research, based in The Netherlands, and sponsored by Tencent, a provider of internet services in China.
The university is working with Tencent to explore breakthroughs that can help feed more people, deliver greater food security and create more food with fewer resources.
The challenge was open to multi-disciplinary teams from across the world, says Dr Silke Hemming, researcher at Wageningen University and Research.
She says: “The teams could be made up of scientists, R&D specialists and students, from different disciplines, such as artificial intelligence, sensor technology and crop physiology.”
The five teams selected to participate in the challenge are each running a 96sq.metre greenhouse compartment, growing cucumbers, between August and December.
A computer and machine learning algorithms or computer models make many decisions concerning production of the crop, with limited human interaction, says Dr Hemming.
“The cucumber has access to nutrients and water via a substrate. There is artificial light, heating and ventilation openings.”
As well as temperature, humidity and CO2 sensors, teams can add any other sensors they think will be useful. Examples include radiation sensors and root zone sensors. Cameras are also being used.
Dr Hemming says: “Teams develop their own algorithm and send this to a digital interface. The computer will decide on the optimum greenhouse climate, based on a wide range of sensors and cameras. Self-learning software will partly take over human decisions.”
However, some tasks, such as wiring and harvesting, will need to be done manually, says Dr Hemming: “Not all robotics are available at the moment.”
The teams will be judged on resources input and crop production output and will have points deducted if their solutions are too complex or expensive.
A control greenhouse compartment is also being run as part of the trial, being operated as a commercial grower would.
All cucumbers are of the same variety, with participants deciding how many plants and how many stems per plant.
Greenhouse production is an efficient way to supply fresh vegetables and fruit with a high content of vitamins and minerals, requiring only a small production area, according to Dr Hemming.
“Production in greenhouses is typically up to 10 times higher than in open fields and uses much less water; 1kg of cucumber requires only five or six litres of water to produce in a greenhouse, whereas 100 times more is required if we do this outside.
“In a closed environment, there is less danger of pests and diseases and less need to use chemicals.”
In the future, more greenhouses will be needed to produce food, says Dr Hemming.
“While the knowledge and experience of a grower is essential, it is difficult to find enough skilled workers in many countries worldwide.
Automation, IT and AI can help the grower access all the information needed and help make complex decisions.
Commenting on the challenge, she says: “We expect to have created a benchmark; a world first. We expect to get new insights into crop physiology and management and develop new co-operation between agriculture and AI.”
Autonomous greenhouses can boost vegetable production and bring greenhouse production to a new level, according to Dr Hemming.
She says: “Autonomous greenhouses can bring further improvement to yield and quality of greenhouse vegetable crops and a substantial reduction of inputs, such as energy, water and crop protection.”
Researchers at Wageningen University and Research are studying different colours of light using LEDs in an attempt to understand how they affect aspects of plant production, such as quality, taste, shelf life and nutritional value.
The research is being carried out in vertical farming laboratories only lit with artificial light. Early findings indicate using more far infra-red light could lead to smaller thinner leaves, with tomato plants diverting more of their energy to fruit production, boosting yield.