UAV Remote Sensing for High-Throughput Phenotyping and for Yield Prediction of <em>Miscanthus</em> by Machine Learning Techniques

• DOI: 10.20944/preprints202206.0120.v1
• Publication Date: 2022-06-09T06:27:06Z
• Authors: Giorgio Impollonia; Michele Croci; Andrea Ferrarini; Jason Brook; Enrico Martani; Henri Blandinières; Andrea Marcone; Danny Awty-Carroll; Chris Ashman; Jason Kam; Andreas Kiesel; Luisa M. Trindade; Mirco Boschetti; John Clifton-Brown; Stefano Amaducci
• Keywords: Interception; Phenomics

Miscanthus holds a great potential in the frame of bioeconomy and yield prediction can help improving logistic supply chain. Breeding programs several countries are attempting to produce high-yielding hybrids better adapted different climates end-uses. Multispectral images acquired from unmanned aerial vehicles (UAVs) Italy UK 2021 2022 were used investigate feasibility high-throughput phenotyping (HTP) novel for crop traits estimation. An intercalibration procedure was performed using simulated data PROSAIL model link vegetation indices (VIs) derived two multispectral sensors. Random forest algorithm estimated with good accuracy (light interception, plant height, green leaf biomass standing biomass) VIs time series predicted peak descriptor 2.3 Mg DM ha-1 RMSE. The study demonstrates UAVs HTP applications providing important information needed increase sustainable production.