UAV-Based High-Throughput Phenotyping for Early Canopy Vigor Quantification in Wheat

Accurate quantification of early wheat growth is essential for improving productivity, resource-use efficiency, and stress adaptation; however, conventional phenotyping approaches remain labor-intensive and difficult to scale for breeding programs. This study evaluated the potential of UAV-based multispectral phenotyping to rapidly and non-destructively estimate early vegetation vigor across diverse wheat genotypes and identify vegetation indices most strongly associated with canopy development. A panel of 221 historical wheat genotypes was evaluated under irrigated field conditions at NIGAB-NARC, Islamabad, Pakistan. Multispectral imagery was acquired at 47 and 61 days after sowing using a MicaSense Altum-PT sensor mounted on a DJI MATRICE 200 RTK V2 platform, producing ultra-high-resolution imagery (1.12-1.50 cm GSD). Eighteen vegetation indices (VIs) were extracted and validated against Canopeo-derived ground cover measurements. Significant variation was observed for all indices, while principal component analysis showed that the first two components explained 92% of the total variation, primarily driven by greenness- and biomass-related traits. Most VIs exhibited strong positive correlations with ground cover, with nine indices showing correlation coefficients greater than 0.85. Among these, GNDVI, CIRE, and RENDVI consistently provided the highest predictive accuracy across both developmental stages, achieving R² values of 0.86-0.91 with RMSE and MAE values below 0.055 and 0.045, respectively. These findings demonstrate that UAV-based multispectral phenotyping provides a rapid, precise, and scalable framework for early-stage wheat monitoring and offers strong potential for integration into high-throughput breeding and precision agriculture programs.