Project
Title
Digital Agricultural Knowledge and
Information System (DAKIS) - Digitale Wissens- und Informationsverarbeitung in der Landwirtschaft - professional
Acronym
DAKIS-pro
Start
01.01.2025
End
30.09.2025
Coordinating Institute
Leibniz-Zentrum für Agrarlandschaftsforschung e.V. (ZALF)
Coordinator
Sonoko Bellingrath-Kimura
Contact
Partner
Deutsches Forschungszentrum für Künstliche Intelligenz GmbH
Hochschule Osnabrück
Forschungszentrum Jülich GmbH
Leibniz-Zentrum für Agrarlandschaftsforschung e.V. (ZALF)
Rheinische Friedrich-Wilhelms-Universität
Bayerische Landesanstalt für Landwirtschaft (LfL)
Hochschule für Nachhaltige Entwicklung Eberswalde
Project team ATB
Allocated to research program
Allocated to research program
Summary
The DAKIS-pro project is developing an automated decision support system (EUS) that integrates ecosystem services (ES) and biodiversity (BioDiv) into agricultural utilisation. It takes site-specific and landscape contexts into account in order to convert yield-orientated agricultural systems into sustainable, productive models. On the basis of data streams and model-based analyses, a user-friendly online EUS is being created that offers farmers, advisors and political stakeholders decision-making options. After the first funding phase, which focussed on the farm level, the second phase will extend the approach to the regional level to promote improved provision of EUS and BioDiv. The challenges lie in particular in scaling up and interdisciplinary cooperation with regional stakeholders.
The ATB is working on the integration of biodiversity indicators into the DAKIS-pro decision support system. The aim is to improve incentive systems for ecosystem services (ESS). In this context, ATB is developing a deep learning-based mapping tool for monitoring biodiversity in permanent grasslands by identifying HNV indicators. The approach also focuses on the development of a generic AI model based on domain adaptation methods, i.e., a detection model suitable for deployment on both aerial (drones) and terrestrial (robots) platforms. Automated monitoring systems combined with sensors, platforms and AI will speed up the biodiversity assessment of species-rich grasslands and the classification of HNV areas, thus facilitating the implementation of results-based payment schemes. Domain adaptation helps to develop more generic AI models with less data and therefore has both scientific and economic significance.
Funding
Bundesministerium für Bildung und Forschung (BMBF)
Funding agency
Projektträger Jülich (PtJ)
Grant agreement number
031B1524H