Predicting Grain Yield in Wheat Using UAV Multispectral and Ground Based Vegetation Indices

High-throughput phenotyping using unmanned aerial vehicle (UAV) multispectral imagery offers a promising approach for predicting wheat yields under variable sowing conditions. This study evaluated the effectiveness of UAV-based vegetation indices compared to the GreenSeeker handheld sensor in estimating yield-related traits in 13 bread wheat genotypes. UAV-based multispectral indices— Normalized Difference Vegetation Index (NDVI), Green Normalized Difference Vegetation Index (GNDVI), Red-edge Normalized Difference Vegetation Index (RNDVI and simple ratio (SR) were captured using a MicaSense sensor at two growth stages [52 and 79 days after sowing (DAS) for timely sown; 16 and 43 DAS for late sown]. Simultaneously, NDVI was recorded using a GreenSeeker handheld sensor for direct comparison with UAV-derived NDVI. UAV-derived indices showed consistently stronger correlations with biological yield (BY), grain yield (GY), and thousand grain weight (TGW), particularly during the anthesis stage. GNDVI and SR emerged as the most predictive indices for BY and GY, while TGW showed stronger associations with early-stage indices. GreenSeeker NDVI correlations were weaker and less consistent across growth stages and sowing conditions. Genotypes such as Phule Samadhan, MACS 2496, and GS 4042 exhibited superior adaptability under late-sown heat stress, maintaining higher vegetation index values throughout. UAV-based multispectral imaging outperformed the handheld sensor in predicting key yield traits and detecting inter-genotypic variation under stress. Statistical and multivariate analyses (ANOVA, PCA, and heatmap visualization) revealed distinct inter-genotypic variability in vegetation indices, effectively distinguishing high-vigor and stress-susceptible wheat genotypes under varying sowing environments. These findings highlight UAV-based multispectral imaging as a robust, efficient, and scalable phenotyping tool for identifying stress-tolerant and high-yielding genotypes, underscoring the importance of phenological timing and optimal index selection in breeding and precision agriculture.