Aerial Insights: Advancing Nitrogen Estimation in Field Crops using Multispectral Imaging

Abstract. Estimating the nitrogen (N) content of crops is crucial for determining key indicators such as nitrogen use efficiency (NUE). Traditionally, most methods for assessing N content have been destructive, time-consuming, and labor-intensive. In this study, we present a non-destructive approach using unmanned aerial vehicle (UAV) multispectral imagery to estimate crop nitrogen content at various growth stages. Multispectral drone data were collected over canola and wheat fields at three growth stages across two experimental sites in Alberta, Canada, over two growing seasons (2023–2024). Simultaneously, leaf tissue samples were gathered from different nitrogen treatment levels, each replicated four times. Multiple machine learning (ML) models were developed and tested to predict plant nitrogen uptake. Our findings indicate that multispectral imagery can estimate N content in canola with a root mean square error (RMSE) ranging from 0.38 to 0.71 and a coefficient of determination (R²) between 0.77 and 0.92. For wheat, the RMSE values ranged from 0.33 to 0.68, with R² values between 0.5 and 0.89. The models showed good transferability across both study sites and two years, suggesting the feasibility of scaling N-content estimation to broader areas. Overall, our results highlight the strong potential of UAV-based multispectral imaging as a reliable, non-invasive tool for estimating nitrogen-related parameters, including plant N-uptake and NUE.