
AI and satellite data for accurate, cost-effective soil analysis
What if we could determine soil texture, carbon content and moisture levels without taking a single soil sample?
Supported by: Food & Bio Cluster Denmark
Project period: June 2026 – December 2026
This question underpins SoilFM, an innovation project exploring whether artificial intelligence and satellite data can deliver more accurate and cost-effective analyses of Danish agricultural soils.
Both Denmark’s Green Tripartite Agreement and forthcoming nitrogen regulations are putting pressure on the agricultural sector to understand soil properties better than ever before.
Farmers and advisers need detailed knowledge of soil variability at field level. However, traditional soil sampling is expensive and only provides measurements at individual sampling points. Through SoilFM, we are exploring whether soil analysis can become more accurate and efficient in the future.
New opportunities with large-scale AI
Recent advances in artificial intelligence have opened up a range of new possibilities in digital soil mapping.
‘It is unclear to what extent this new technology can be applied to Danish soil conditions, and that is what we will be investigating in SoilFM,’ says Mikkel Bue Lykkegaard, PhD and Data Science Lead at the Danish Technological Institute.
Towards a scalable solution
The Danish Technological Institute is leading the technical work, testing whether ‘geospatial foundation models’ – large-scale AI models trained on satellite data – can predict a range of soil properties, including soil texture and organic carbon content.
This proof-of-principle project focuses on reducing technical, data-related and organisational uncertainties associated with the method.
The project will deliver a report on the findings, a prototype model and a plan for further developing and testing the technology on a larger scale.
SoilFM is being carried out in collaboration with climate tech company PEATZ.
Contact us
Would you like to learn more about SoilFM?
Contact Mikkel Lykkegaard
PhD and Data Science Lead, Danish Technological Institute
mbly@teknologisk.dk
+45 72203247