Autonomous Greenhouse Challenge 4th Edition

Earlier editions of the Autonomous Greenhouse Challenge have been carried out at Wageningen University & Research, the Netherlands, during 2018-2022, with teams growing cucumbers, cherry tomatoes, and lettuce crops autonomously. During these editions we have shown that computer algorithms can increase greenhouse crop production, save energy, and yield higher net profits. We have demonstrated that artificial intelligence could be superior to human intelligence, and could potentially control indoor farming in the future.

Wageningen University & Research works on autonomous greenhouses, climate control, crop modelling, use of sensors, data, artificial intelligence, computer vision, and robotics. We develop new technology and new algorithms for and together with industry. These technologies facilitate growers to produce higher quality, healthy crops in a sustainable way and using less resources.

The Autonomous Greenhouse Challenge is sponsored by Tencent and David Wallerstein (CXO Tencent). The Challenge is open to national and international teams.

The Challenge set-up

Online challenge and Hackathon event (April to June 2024)

Teams performed different tasks in an online challenge and Hackathon event. Teams showed their skills in computer vision using image analysis algorithms to estimate plant traits on provided pictures (part A); in machine learning by developing algorithms to grow dwarf tomatoes virtually using the WUR greenhouse and crop models (part B); in few-shot computer vision bonus task to estimate insects on a limited number of yellow sticky trap pictures (part C); in pitching their approach in front of our international jury (part D). The top 5 teams for the greenhouse growing experiment have been selected!

Greenhouse growing experiment and International Autonomous Greenhouses Event (August 2024 to January 2025)

The selected top 5 teams compete against each other and a group of growers who will operate a reference greenhouse compartment. Each team gets their own compartment at WUR’s greenhouse facility in Bleiswijk, the Netherlands. We’ll provide the plants, greenhouse, data platform, standard sensors, and cameras. You develop the machine learning and computer vision algorithms needed to grow the dwarf tomatoes remotely and autonomously. Goal: Selection of winner of the 4^th edition of the Autonomous Greenhouse Challenge.

Jury

In-Bok Lee | Full Professor at Soul National University, Laboratory of Aero-Environmental & Energy Engineering (A3EL).

Expertise in environmental controls of smart farms (greenhouses, livestock) and residential buildings as well as urban agriculture, energy saving technologies and renewable energy, ICT applications including virtual reality, digital twin and machine learning.

Kathy Steppe | Full professor at Ghent University, Faculty of Bioscience Engineering, Belgium, steering committee member of UGent Agrotopia.

Expertise in plant ecophysiology, plant-water relations and carbon metabolism, plant sensors for monitoring and stress detection, plant microCT- and
PET-scans, process-based and functional-structural plant modelling, and development of advanced plant-based control strategies.

Leo Marcelis | Full professor Horticulture and Product Physiology at Wageningen University, The Netherlands.

Expertise in experimental and simulation research on growth and functioning of crops in greenhouses and vertical farms, physiology, growth and product formation of plants and plant organs, improving crop production and quality including energy saving, efficient use of (LED)light, water and nutrients.

Sponsors