AI in
Soil Science

Describing soil profiles in the field is fundamental to soil protection, but it is time-consuming and demanding. The high heterogeneity of soils and the resulting large number of soil horizons also mean that the same soil profile is often described differently by different soil scientists.

To simplify soil profile description and improve the quality of soil science datasets, the Federal Institute for Geosciences and Natural Resources (BGR) and the Berlin University of Applied Sciences and Technology (BHT) have jointly developed an AI-based solution for the automated segmentation and classification of soil horizons in profile images. The model is available through a web application that assists with annotation.

At the workshop, we want to bring soil scientists and interested participants together, present the application, and discuss the future of digital soil protection. Participants can take part in a short interactive study in which soil profiles are annotated with and without AI assistance. The results will be presented at the end of the workshop.

A poster session will accompany the workshop. We warmly invite all participants to present their own projects, research approaches, or practical examples on “Digital methods in soil protection and soil science”.