AI-based integration of remote sensing and citizen science data to infer biodiversity in forests

Despite the increasing threat to essential forest ecosystem services due to the biodiversity and climate crises, there is no robust biodiversity monitoring. The main reasons for this are the lack of accessibility and heterogeneous formats of valuable data and suitable analysis methods. The aim of the iForest project is to bring together the entire available repertoire of geoscience and citizen science data (drones, Flora Incognita, iNaturalist) with remote sensing data in order to use them to map the biodiversity of forest communities using artificial intelligence (AI) methods. This is to be achieved by using common development environments for AI methods, which, among other things, enable the flexible integration of extremely high data volumes. The use of heterogeneous data, e.g. from authorities and citizen science, should be made possible. In addition, the AI methods should also detect non-linear correlations in high-dimensional data sets.

In the first phase, the project pursues the in-depth development of a concept and the preparation of an application for a second phase, which will focus on the implementation of the aforementioned objectives. In addition, initial scientific work will be carried out in the project to allow an effective and rapid entry into a possible second phase.

Project lead: Dr. Daniel Doktor

Helmholtz-Centre for Environmental Research

Das Projekt im Interview (Content in German)

iForest: Wie KI-basierte Biodiversitätsforschung Wälder schützen kann

Um Biodiversitätsrisiken abschätzen und passende Schutzmaßnahmen entwickeln zu können, sind belastbare Daten zum Zustand eines Ökosystems und dessen Veränderung unverzichtbar. Dr. Daniel Doktor vom Forschungsprojekt iForest will die Biodiversität von Wäldern mithilfe von Künstlicher Intelligenz (KI) abbilden und dazu Geo-, Citizen-Science- und Fernerkundungsdaten zusammenführen.

Zum Interview auf der FONA-Seite.