Latest from the project

Graphical overview of the drone phenotyping experiment

With the drone-phenotyping experiment, we collected high-resolution image data on barley, durum and bread wheat trial fields. In each field, multiple genotypes were sowed. We used three cameras (RGB, multi-spectral, thermal) and performed 10 flights over each field to cover the whole growing season (from February to June).

This produced a lot of image data that will be used for:

1. the calculation of vegetation indices (e.g. NDVI: normalized difference vegetation index; complete list here)
2. the development of DNN (deep neural network models) for phenotypic predictions: i) can we use image data to predict yield? ii) can we use early image data to predict future growth? iii) can we use lower-resolution data to obtain predictions of comparable accuracy?
3. genomic predictions, when coupled with genotyping/sequencing data

The figure below gives a visual overview of the drone-phenotyping experiment:

Figure: Graphical overview of the drone-phenotyping experiment

The drone phenotyping experiment ended successfully on June 10 2024 at SIS facilities in S. Lazzaro di Savena (BO), and on June 12 at CREA-GB in Fiorenzuola (PC).