CAgriLab is developing a shared digital environment to make data, experiments and knowledge from Agricultural Living Labs easier to discover, compare and reuse.
Follow the project's progress from initial requirements and data collection to harmonisation, Digital Twins, AI tools and natural-language search.
Key technical and research achievements from across the project.
The project began by defining the needs of Living Labs and farmers, including how regenerative agriculture practices and experiments could be represented, compared and discovered. Early work identified the importance of contextual information such as soil, weather, management practices and location when interpreting agricultural evidence.
Project partners began collecting real-world data to support development and testing of CAgriLab. HAMK collected soil samples, moisture measurements, field imagery, weather data and agricultural sensor data, while CGFP prepared field and farm-management data for integration with the platform.
The first CAgriLab platform infrastructure was established using Pontus-X. Initial functionality enabled data to be uploaded and downloaded, while work began on compute-to-data capabilities.
CAgriLab partners developed tools for harmonising agricultural data from different sources, using GLOSIS as a foundation for soil data. HAMK also completed a field data-collection season covering forage, soil and biodiversity measurements.
The CAgriLab Pontus-X portal became operational and was successfully tested. Initial functionality allowed data and algorithms to be published and accessed through the platform.
CAgriLab introduced functionality allowing users to publish data without requiring a blockchain wallet, reducing a significant usability barrier. The project's first system architecture had also been drafted.
HAMK progressed several AI and computer-vision applications designed to support Living Lab data collection, including image-based soil structure assessment, biodiversity measurement and estimation of grass and clover composition.
Prototype algorithms were developed for estimating grass/clover ratios, above-ground forage biomass in multi-species systems and detecting cow activity.
A prototype natural-language search capability was developed, alongside functionality for uploading and visualising Digital Twins. Wallet-free publishing and data harmonisation tools had progressed to production status.
Development moved towards deeper integration of the project's components, including the Digital Twin viewer and builder and a container-based computing environment for running algorithms alongside the data infrastructure.
Explore the tools, Living Lab data and digital infrastructure being developed through CAgriLab.
