Assessing nitrogen content in flood-irrigated rice plantations using uav-based multispectral imageries

Abstract

Nitrogen is essential for agricultural crops, especially rice, and monitoring its demand is crucial. Remote sensing offers a fast and efficient alternative to field sampling, facilitating fertilizer management. This study estimated agronomic parameters related to nitrogen status in flooded rice, total above-ground biomass (AGB), leaf nitrogen content (LNC), leaf area index (LAI), and productivity, using a multispectral UAV sensor and a proximal optical reflectance sensor (Crop Circle), compared to traditional sampling in experimental plots. Simple and multiple linear regressions were applied. Field sensor data achieved coefficients of determination (R²) of 0.89 and 0.85 for LNC and LAI, respectively. The growing stage significantly affected the performance of vegetation indices, whereas rice variety had no significant impact. Model performance varied with phenological stage, with AGB, LNC, and LAI best estimated during the reproductive stage, and yield during grainfilling. Effective results were also obtained when combining all stages. UAV-based analysis proved a promising alternative, overcoming satellite cloud cover limitations, offering high cartographic accuracy, historical records over larger areas, and shorter operational time compared to field optical sensors and traditional methods.