Developing and identifying high-yielding crop varieties adapted to specific environmental conditions requires substantial effort to evaluate phenotypic traits such as LAI, plant height, biomass, and yield. These traits play a crucial role in ensuring stable productivity and efficient resource use. Traditional methods for assessing crop phenotypes rely on destructive field sampling and hand-held instrument measurements, which are time-consuming and often lack representativeness. In contrast, remote sensing offers an innovative, rapid, non-invasive, and efficient approach to quantifying the structural and functional traits of field crops in a timely manner.
The goal is to predict basic crop traits from full spectral responses across the solar spectrum, collected with a hyperspectral camera. Multiple wheat cultivars by CIMMYT were measured across different years and phenotypic stages. You are encouraged to compare classic remote sensing approaches (i.e., calculate spectral indices from the spectral curve and assess their correlation with traits) as well as a “feature-less” machine learning approach (i.e., use the entire spectral curve to make predictions without creating any indices, perhaps with neural networks).
Spectral indices - https://awesome-ee-spectral-indices.readthedocs.io/en/latest/index.html
Crop remote sensing - Araus, J.L., Kefauver, S.C., Vergara-Díaz, O., Gracia-Romero, A., Rezzouk, F.Z., Segarra, J., Buchaillot, M.L., Chang-Espino, M., Vatter, T., Sanchez-Bragado, R., Fernandez-Gallego, J.A., Serret, M.D., and Bort, J. (2022). Crop phenotyping in a context of global change: What to measure and how to do it. J. Integr. Plant Biol. 64: 592–618.