Estimating buckwheat yield based on sowing methods with UAV-RGB imagery and time-series data

Drill seeding and broadcast seeding are two common methodologies employed for crop establishment in agricultural practices. While as UAV-based remote sensing and vegetation indices serve as valuable tools for crop monitoring and precision management in modern precision agriculture. This study aims to evaluate the effects of broadcast seeding and drill seeding on the growth characteristics and yield components of common buckwheat. Utilizing UAV-based RGB imagery and time-series data, the research compares seed efficiency between methods by assessing seed quantities required for optimal emergence and growth stability. Additionally, it investigates how row spacing in drill seeding influences plant density and potential yield outcomes. Vegetation monitoring employed three cost-effective indices—Excess Green Index (EGI), Normalized Green–Red Difference Index (NGRDI), and Green Leaf Index (GLI)—derived from RGB images to assess plant health and biomass accumulation. Results indicated that while parameters such as fresh weight, seed weight, stem diameter, number of branches, lodging rating, and seeds per plant showed no significant differences across sowing methods, plant height was notably affected (p = 0.0493), suggesting row spacing impacts growth traits. UAV imagery revealed that average plant height estimates were lower than ground measurements, consistent with prior research, with high-resolution images enhancing estimation accuracy. The study highlights the potential of RGB-based remote sensing as a cost-effective and practical tool for monitoring crop phenology, particularly plant height, which correlates strongly with yield. Findings suggest that optimized UAV imaging can reliably evaluate plant growth traits and assist in selecting the most effective sowing strategies for buckwheat production.