New Dutch project for detecting plant diseases using AI
The sustainability project Remote Sensing for Floriculture (RS4F2), was launched on Friday, January 24, at the Dutch Landgoed Tespelduyn in the presence of all involved partners. This project focuses on detecting plant diseases and crop abnormalities using artificial intelligence (AI), drones, satellites, and other sensors.
Last year, researchers from the Dutch western Dune and Bulb region (Duin- en Bollenstreek) successfully developed an AI model that enables drones to detect plant diseases. This yielded promising results. With the launch of this highly anticipated follow-up, organisations such as Unmanned Valley, Greenport Duin- en Bollenstreek, NL Space Campus, Economic Board Duin- en Bollenstreek, Holland Rijnland, and others are once again joining forces to accelerate the sustainability of floriculture.
Sustainability through innovation
This new project sets the bar even higher. In addition to drones, researchers will now also utilise satellite imagery, ground sensors, and camera footage to detect plant diseases. By combining various technologies, the expectation is that it will be possible to accurately identify which plants require protection and, more importantly, which do not. The ultimate goal is for growers to seamlessly integrate this technology into their operations, allowing for earlier disease detection and the more efficient, sustainable, and precise use of crop protection agents. This could potentially result in significant cost savings and reduce the risk of crop losses.
New drones
Currently, drones fly relatively low and slow over fields to ensure high-quality image collection. These images must be precise down to the millimeter. In the next phase of the project, researchers will explore ways to speed up this process so that larger areas can be mapped. By employing new data acquisition techniques and integrating these results with other datasets, the research team hopes to collect information more quickly and efficiently.
Additionally, the use of so-called “drone boxes” is being explored. These automated systems allow drones to conduct pre-planned flights without a pilot. The drones are on standby in a protected “box” near the fields and can take off at any given moment for real-time data collection.
“The first phase took us further and provided more insights than we had dared to hope, thanks in part to the collaboration with all stakeholders. I believe that technological knowledge in floriculture, with its complex business and sustainability challenges, can contribute not only to solutions but also to new revenue models for the region,” says Theo de Vries, director of Unmanned Valley.
Data from space
The quality of satellite imagery has significantly improved in recent years. This offers the potential for large-scale insights that could play a crucial role in this type of precision agriculture. Furthermore, satellites provide essential data on environmental factors such as weather conditions and soil moisture. Satellite data opens up opportunities to accelerate the practical implementation of this project’s findings by growers.
Expansion
The new research is set to expand significantly compared to the first project. In addition to drone and satellite data, researchers will investigate other efficient methods for gathering relevant AI data. This includes cameras on tractors, agricultural robots, and ground sensors. Combining these data sources could ultimately make it possible to predict the emergence and spread of plant diseases.
While the initial focus was on detecting Botrytis in tulips and hyacinths, the scope is now being broadened. Researchers are examining which other crops and diseases the model could be adapted for, making the technology widely applicable both within and beyond the floriculture sector.
The agriculture of the future
The upcoming year will focus on collecting new data. Researchers will spend a lot of time in the fields as soon as spring arrives. The newly gathered data will be used to further refine the model. Additionally, efforts will be made to ensure practical applicability, enabling growers to easily integrate these models with their existing equipment. The RS4F team expects to present its findings around December 2025.
Join 17,000+ subscribers
Subscribe to our newsletter to stay updated about all the need-to-know content in the agricultural sector, two times a week.
