Gastvortrag: Dr. Brian N. Bailey (University of California, CA, USA)

Titel: "Leveraging 3D plant modeling to enable holistic high-throughput phenotyping of plant form and function in the field"

Zusammenfassung: High-throughput phenotyping (HTP) has advanced rapidly, yet key limitations persist: field-based end-to-end pipelines remain fragmented and difficult to integrate, and extracted traits are often limited to visible features from computer vision (e.g., fruit counting) or non-specific spectral indices that indicate “stress” but lack diagnostic power. We address these challenges through the GEMINI project, developing an open-source, end-to-end HTP pipeline and leveraging 3D crop models to extract holistic traits indicative of interpretable plant function. Central to our approach is Helios, a 3D plant systems modeling framework used to generate large volumes of auto-annotated synthetic data for training AI models to understand plant biophysics and estimate traits from multimodal imagery. We further train AI models to estimate biophysical model parameters from sensor data, enabling 3D model-based prediction of plant behavior under novel environmental conditions. This integration of mechanistic modeling with data-driven AI offers a path toward more robust, interpretable, and transferable phenotyping of plant form and function in field settings.

Über Dr. Brian N. Bailey
seit 2016 Associate Professor im Department of Plant Sciences, University of California, Davis
2015 Entwicklung verbesserter biophysikalischer Modelle zur Ausbreitung des Traubenmehltaus; USDA-ARS in Corvallis, OR
2015 Ph.D. in Mechanical Engineering, University of Utah
2011 B.S./M.S. Mechanical Engineering, University of Utah

Wann?
Donnerstag, den 16. März 2026, 11:30-12:30 Uhr

Wo?
Leibniz-Institut für Gemüse- und Zierpflanzenbau (IGZ), Theodor-Echtermeyer-Weg 1, 14979 Großbeeren
Hörsaal (Verwaltungsgebäude)

Vortragsprache: Englisch

Anmeldung
Wir bitten um Anmeldung unter lectures@igzev.de.