Abstract
Optimal crop management is necessary for sustainable food and feed production. Continuous monitoring of crop canopies is important in order to adapt management processes, e.g. fertilisation and plant protection, but also time-consuming. This task can be supported using remote sensing data. Many previous studies applied vegetation indices in order to retrieve information on crop canopies. They can be calculated easily and provide information on relative differences in crop canopies regarding e.g. plant health and biomass production. Vegetation indices, however, require continuous calibration as well as validation and often do not use the full spectral resolution of many sensor systems. An alternative to vegetation indices are radiative transfer models. These models describe the interaction of solar radiation and vegetation canopy. Radiative transfer models have low calibration as well as validation needs and allow the use of all available spectral information. In this study, a dataset was simulated based on the radiative transfer model PROSAIL. An artificial neural network was trained and tested with this dataset. For further evaluation, field experimental data of autumn and spring sown wheat with different nitrogen fertilisation levels at the experimental farm Groß-Enzersdorf of the University of Natural Resources and Life Sciences Vienna (BOKU) was used. Preliminary canopy parameter estimations on LAI and chlorophyll content were promising. In future, this work can help monitor crop canopies, optimise management processes and improve the sustainability of food as well as feed production. © ECPLF 2022. All rights reserved.