In January 2025, the Remote Sensing for Floriculture (RS4F) project officially launched at the Tespelduyn Estate in Noordwijkerhout. This project builds on earlier research (Remote Sensing for Floriculture 1), which led to developing an AI model capable of detecting plant diseases, such as botrytis, at an early stage using drone imagery. This technological innovation received widespread media attention in 2024 and won a Computable Award.
The technology is expanding. In addition to drone imagery, satellite images, soil sensor data, and camera footage are now being used to improve the accuracy of disease detection. By combining these various data sources, we aim to create a more powerful and versatile model.
During the spring, the RS4F team actively collected new data. Our pilots flew over the blooming fields weekly to capture detailed aerial images. These images form the foundation for training the AI model, which is increasingly adept at recognising early signs of disease in plants.
After image collection, the intensive process of annotation begins. Each image is carefully analysed and labelled with information about the health status of the plants. This meticulous work is crucial to improve the AI’saccuracy of AI. “We’ve already put in many hours. But if we want to apply precision agriculture, we need to be precise,” says Walter Kort, RS4F Operations Officer.
The focus will be further annotating the newly collected data in the coming months. In parallel, we are exploring whether drone data can be linked to satellite imagery, making the technology more scalable. Another research direction investigates whether it’s possible to develop a more objective predictive capability by combining data from a dedicated field-based weather and soil monitoring station with additional satellite data.
The project will run until the end of the year and is supported by the NL Space Campus, Unmanned Valley, Greenport Duin-en Bollenstreek, Economic Board Duin- en Bollenstreek and Holland Rijnland.
We’ll be sharing more updates on our website and LinkedIn page soon.