Content area
Unmanned aerial vehicles (UAVs) equipped with multi-sensor remote sensing technologies provide an efficient approach for mapping spatial and temporal variations in vegetation traits, enabling advancements in precision monitoring and modeling. This study’s objective was to analyze UAV multiple sensors’ performance in monitoring winter wheat chlorophyll content (SPAD), plant nitrogen accumulation (PNA), and N nutrition index (NNI). A two-year field experiment with five N fertilizer treatments was carried out. The color indices (CIs, from RGB sensors), vegetation indices (VIs, from multispectral sensors), and temperature indices (TIs, from thermal sensors) were derived from the collected images. XGBoost (extreme gradient boosting) was applied to develop the models, using 2021 data for training and 2022 data for testing. The excess green minus excess red index, red green ratio index, and hue (from CIs), and green normalized difference vegetation index, normalized difference red-edge index, and normalized difference vegetation index (from VIs), showed high correlations with three N indicators. At the pre-heading stage, the best performing CIs correlated better than the VIs; this was reversed in the post-heading stage. CIs outperformed VIs in SPAD (CIs: R2(coefficient of determination) = 0.66, VIs: R2 = 0.61), PNA (CIs: R2 = 0.68, VIs: R2 = 0.64), and NNI (CIs: R2 = 0.64, VIs: R2 = 0.60) in the pre-heading stage, whereas VI-based models achieved slightly higher accuracies in post-heading and all stages compared to CIs. Models built with CIs + VIs significantly improved the models’ performance compared to single-sensor models. Adding TIs to CIs and CIs + VIs further improved the models’ performance slightly, especially at the post-heading stage, resulting in the best model performance with three sensors. These findings highlight the effectiveness of UAV systems in estimating wheat N and establish a framework for integrating RGB, multispectral, and thermal sensors to enhance model accuracy in precision vegetation monitoring.
Details
Vegetation;
Nutrition assessment;
Wheat;
Accuracy;
Sensors;
Nitrogen;
Corn;
Remote sensing;
Data processing;
Remote monitoring;
Unmanned aerial vehicles;
Data integration;
Temperature;
Temporal variations;
System effectiveness;
Winter wheat;
Chlorophyll;
Normalized difference vegetative index;
Multisensor fusion
; Wang, Wentao 1 ; Liu, Suyi 1 ; Hui, Xin 1 ; Zhang, Haohui 1 ; Yan, Haijun 2 ; Maes, Wouter H 3
1 College of Water Resources and Civil Engineering, China Agricultural University, Beijing 100083, China;
2 College of Water Resources and Civil Engineering, China Agricultural University, Beijing 100083, China;
3 UAV Research Centre, Department of Plants and Crops, Ghent University, 9000 Ghent, Belgium;