Breeding alfalfa (Medicago sativa L.) presents challenges due to its perennial characteristics. Its biomass yield needs to be characterized several times annually. However, conventional techniques in characterizing plants are a bottleneck in plant breeding. This study explores the potential of UAS-based sensors for phenotyping to estimate alfalfa biomass, plant height, and forage quality. Plant height is considered a surrogate for biomass yield. We correlated ground-based measurements with UAS-based spectral and structural data from RGB, multispectral, and LiDAR sensors. Integrating RGB, multispectral, and plant height data yielded more accurate biomass predictions. Machine learning models showed that combining RGB and LiDAR data improved plant height predictions. For forage quality, vegetation indices from the dual multispectral sensor showed strong correlations with ADF and NDF, while single bands from the sensor are highly correlated with CP. These findings highlight the potential of UAS-based HTP for efficiently and accurately characterizing alfalfa traits.
