Registration for the 4th edition of the International Autonomous Greenhouse Challenge starts 1 November. This edition, Wageningen University & Research (WUR) and co-developer Tencent challenge teams to remotely grow dwarf tomatoes using artificial intelligence (AI).
The competition starts in spring 2024 and consists of an Online Challenge in which AI and horticultural experts grow a virtual crop. The top five teams from the Online Challenge will get their own greenhouse compartment at WUR in Bleiswijk, the Netherlands, in autumn 2024. There they will compete with each other to grow real dwarf tomatoes in the most sustainable and autonomous way possible.
With an ever-growing world population, the demand for fresh and healthy fruits and vegetables is increasing. According to the Food and Agriculture Organization (FAO) of the United Nations, it will require 60% greater food production to feed a global population estimated to reach 9 billion by 2050. Autonomous greenhouses can ensure that more people are fed with vitamin- and mineral-rich produce. In addition, these techniques can help improve food safety and increase production of healthy vegetables while using fewer resources such as water and energy.
The potential of autonomous growing has been successfully demonstrated in previous editions of the challenge. This year, dwarf tomatoes take center stage. Tomatoes are traditionally grown in a high-wire cultivation system and they grow meters tall, requiring a lot of manual labor. Dwarf tomatoes, on the other hand, remain compact, only about 40 cm high. This makes them ideal for full automation and cultivation in urban environments.
Challenge as a connection between AI and food production
By organising the Autonomous Greenhouse Challenge, WUR and co-developer and technology company Tencent want to further connect the world of AI and food production, develop knowledge and make this knowledge publicly accessible. The Challenge is thus a contribution to global efforts to make food systems more sustainable.
Online challenge for AI experts
The first part of the 4th edition of the Autonomous Greenhouse Challenge will take place from 1 April to 1 June as an Online Challenge. This part of the Challenge is focused on testing participants' machine learning and computer vision skills.
During the computer vision part of the challenge, teams must develop algorithms based on training images of dwarf tomato plants. Those autonomous algorithms must assess plant characteristics, for example, plant height, leaf area, the number of fruits and their maturity.
In the machine learning part, teams are challenged to use machine learning to develop algorithms to autonomously control ventilation, heating and lighting in a virtual greenhouse, among other things. The algorithms should stimulate the growth of a virtual dwarf tomato crop while minimising the use of energy.
Hackathon event for participating teams
All participating teams will then be invited to a Hackathon event on 6/7 June at WUR in Bleiswijk. During this event, the points for the two parts of the Online Challenge will be announced. Teams can score extra points by performing an additional task and by giving a pitch of their approach in front of an international jury. The five best teams with the most points from all parts will be selected during the Hackathon event. They will receive their own greenhouse section in the autumn to prove their skills in a real dwarf tomato crop.
Who can participate?
Teams with a minimum of 3 members. Teams must consist of at least one horticulture and one AI expert. At least one team member must be a student. A participant can only be part of one team and register once. The Challenge welcomes teams from different countries and continents, and the organisers encourage collaboration between experts from start-ups/companies with students and researchers from universities/research centres. Participants from previous editions may also participate: we would love to see teams return to push their boundaries and develop new knowledge.