Accurately predicting rice yield and its components is crucial for optimizing agricultural practices and ensuring food security. Traditional methods of assessing crop status wwcan be time-consuming and labor-intensive. This study investigated the use of drone-based multispectral imagery and machine learning to improve the prediction of rice yield and yield components. Time-series VIs were collected from 152 rice samples across various nitrogen treatments, transplanting times, and rice varieties in 2023 and 2024, using an UAV at approximately 3-day intervals. A four-parameter log-normal model was applied to analyze the VI curves, effectively quantifying the maximum value, spread, and baseline of each index, revealing the dynamic influence of nitrogen and transplanting timing on crop growth. Machine learning regression models were then used to predict yield and yield components using the log-normal parameters and individual VIs as input. Results showed that the maximum (a) and variance (c) parameters of the log-normal model, derived from the VI curves, were strongly correlated with yield, grain number, and panicle number, emphasizing the importance of mid-to-late growth stages. Among the tested VIs, NDRE, LCI, and NDVI demonstrated the highest accuracy in predicting yield and key yield components. This study demonstrates that integrating log-normal modeling of time-series multispectral data with machine learning provides a powerful and efficient approach for precision agriculture, enabling more accurate and timely assessments of rice yield and its contributing factors.

Purpose Rice (Oryza sativa L.) production is threatened by bacterial leaf blight (BLB), particularly from the virulent K3a strain. Reliableand efficient phenotyping is essential for breeding resistant cultivars, but traditional manual methods are time-consuming and prone to subjectivity. Methods This study assessed the accuracy of smartphone-based image analysis for measuring lesion length (LL) and lesion ratio (LR) in BLB-infectedrice. A total of 102 recombinant inbred lines (RILs) derived from Chamdongjin and Sindongjin were evaluated using both manual methods and imageanalysis. Images were captured in the field using a smartphone (12MP, 4000×3000 pixels), without any auxiliary equipment. The images were analyzedusing ImageJ and Python-based tools and compared with manual measurements. Results Image-based data showed strong correlation with manual results(r = 0.84–1.00), and principal component analysis (PCA) explained 94.14% of the variance, effectively distinguishing resistant and susceptible lines.Statistically significant differences were observed between groups (p < 0.01), confirming reliability. Images taken at 3x magnification showed highaccuracy (r = 0.88–0.92), while those at 1x magnification showed reduced accuracy (r = 0.76–0.79), likely due to the inclusion of surrounding plants—highlighting the importance of proper field design and image framing. Conclusion This study suggests that smartphone-acquired images, even withoutauxiliary equipment, may serve as a feasible and sufficiently accurate source for image-based phenotyping. This accessible approach has the potential tosupport rapid, low-cost data collection in field settings, which could contribute to breeding and disease resistance research.