UAV image analysis for detecting rice seedling gaps and gap effect on grain yield

• A novel approach utilizing UAV images to detect rice seedling gaps. • Algorithms were developed to extract plots, row lines, and seedling gaps from UAV images. • Cultivars explained 53.3 % the yield variability. • On average, a 1 % increase in seedling gap fraction led to a ∼0.974 % reduction in rice yield The spatial distribution of seedlings influences the structure and productivity of rice by impacting canopy light distribution, radiation use efficiency, and ultimately crop yield. This study proposes a novel approach utilizing high-resolution UAV images to detect seedling gaps, providing an alternative method to estimate seedling establishment in densely planted crops such as rice. Image processing algorithms were developed to extract plot and row lines from UAV images, enabling the estimation of gaps between rice seedlings within rows. Analyses revealed that cultivar differences explained the largest proportion of yield variability (53.3 %), followed by the year × cultivar interaction (24.2 %), year (9.8 %) and gap fraction (1.9 %). On average, a 1 % increase in seedling gap fraction led to a 0.974 % reduction in rice yield. The impact of seedling gaps on yield was not uniform and varied with the size of the gaps and the stage of crop development at the time of measurement. For instance, yield was predicted to decrease when gap fraction was above 4.5 % at 300°D > 10 °C, while even a smaller gap fraction (>0.5 %) led to yield reductions at 600°D. This study highlights the importance of monitoring seedling gaps to guide adaptive crop management, offering insights to optimize decisions on replanting or adjusting planting strategies to minimize gaps and improve yield.