Improving crop disease detection through multi-view drone imagery

Drones are increasingly used to detect crop diseases from the air. However, they typically photograph plants from directly above and may therefore miss the upright and lower leaves on which many diseases develop.

This DFG-funded project, led by Dr. Rene H. J. Heim together with doctoral researcher Davide Mattioli, investigates whether tilting a drone-mounted camera and observing plants from multiple viewing angles can make disease-detection models more robust and reliable.

Using Cercospora leaf spot in sugar beet as a model system, the team collects multispectral data over several growing seasons and across all stages of crop development. The project examines whether multi-view imagery improves the reliability of disease-detection models and whether particular viewing angles are more or less informative for crops with upright leaves compared with crops whose leaves lie relatively flat.

The findings could support more targeted fungicide applications and contribute to the breeding and evaluation of more disease-resistant crop varieties.

Principal Investigator

Dr. Rene H. J. Heim

Doctoral researcher

Davide Mattioli

Duration

Until 30 April 2029

Host institution

University of Göttingen, in collaboration with the Institute of Sugar Beet Research (IfZ)

Funded by the Deutsche Forschungsgemeinschaft (DFG, German Research Foundation): DFG project 521313940.