WUR-AgriFoRwArdS joint summer school

Digital plant phenotyping and agri-food robotics

Not yet scheduled (last in July 2024)

Scope

At Wageningen University and Research and at the centre for doctoral training AgriForwards in the UK, research is done on the topic of plant phenotyping and plant monitoring using techniques from computer vision and robotics. In this summer school, we want to bring together young researchers from both institutes with the aim to share knowledge, learn together, and to build an international research community.

AgriForwards is a centre for doctoral training in agri-food robotics. It is a joint program with 50 PhD candidates from the universities of Lincoln, Cambridge and East-Anglia. The PhD candidates work on diverse topics, related to sensing and perception, mobile autonomy, robotic manipulation, and human-robot collaboration, with applications in the agri-food sector.

A summer school was held in July 2024, in the form of a hackathon with the concept of learn-by-doing. PhD candidates, postdocs, and young researchers from both institutes were mixed into small teams to actively work together on a research challenge of your preference, ranging from implementing AI methods to process images of plants and fields to programming robots to operate in agricultural environments. The teams were supervised by experts from WUR and AgriFoRwArdS.

The summer school was very hands-on, meaning that a lot of time was spent programming, working with data, and testing solutions. It was an immersive experience, with all staying in a hotel, working at the hotel or in one of the WUR labs. A range of social activities took place to get to know each other and to have fun together.

Course set-up

The course consisted of a set of motivational lectures and group work where participants work on a agri-food robotics challenge. The course started off with a poster carousel in which participants introduced themselves. Posters remained in the lecture room throughout the course. 

Most of the days were spent working on the challenges. They were announced on the website and participants selected beforehand which one they wanted to work on. To tackle the challenge participants developed concepts, wrote code, tested their system, and evaluated the results. Apart from the technological advancements, participants were asked to reflect on the sustainability of their approach and on the ethical, legal, and social aspects of the technologies developed. On the last day of the summer school, results from all groups were shared and discussed. 

The challenges
3 types:

• Database challenges: some challenges were focused on plant phenotyping, with the aim to extract information about plant traits from images and 3D point clouds. Other challenges contained aerial images collected by UAVs to monitor plant development or detect diseases.
• Simulation challenges: robotic simulations were made available to develop methods for robotic perception and control. Examples: a robot arm with camera as end effector that needed to actively reconstruct simulated plants, and a simulated vineyard with the aim to reconstruct the environment or before autonomous navigation.
• Robot challenges: some real robotic setups were available.

Course Organisers in 2024

• Gert Kootstra, Wageningen University and Research
• Grzegorz Cielniak, University of Lincoln
• Marc Hanheide, University of Lincoln
• Kate Smith, University of Lincoln
• April Cadzow, University of Lincoln
• Claudius van de Vijver & Miriam van Heist, PE&RC  

More information

Claudius van de Vijver (PE&RC)
Email: claudius.vandevijver@wur.nl