invaPICS Phenology of Invasive Plants from Citizen Science Images

16.11.2026

The invaPICS project, which is based at Justus Liebig University Giessen, uses artificial intelligence and citizen science data to investigate the phenology of invasive plant species. Knowing when plants flower, produce fruit and undergo other seasonal changes is vital for studying how they respond to environmental conditions and the ecological processes associated with biological invasions. Our research combines large-scale biodiversity observations from citizen science platforms with cutting-edge deep learning models to automatically identify plant phenological stages in photographs. By analysing extensive image datasets across species, geographic regions and environmental conditions, we aim to improve our understanding of plant phenology on a broad spatial scale. The computational infrastructure provided by de.NBI Cloud is crucial in enabling this research. Access to GPU resources enables us to process large image collections efficiently, extract visual features, and train and evaluate deep learning models. We can then apply these models to new datasets. These computational capabilities are particularly valuable when working with modern computer vision architectures and large volumes of biodiversity data. Through the integration of citizen science, artificial intelligence, and cloud computing, invaPICS seeks to develop scalable methods for extracting ecological data from biodiversity images. These methods have the potential to support not only invasive species research but also broader applications in biodiversity monitoring, ecological research, and the study of global environmental change.

Yves Philippe Klinger
Yves Philippe Klinger from Landschaftsökologie und Landschaftsplanung, Institut für Landschaftsökologie und Ressourcenmanagement - Justus-Liebig-Universität Gießen​