Winter Wheat Nitrogen Estimation Based on Ground-Level and UAV-Mounted Sensors.

Area of Science:

• Agronomy
• Remote Sensing
• Plant Physiology

Background:

• Optimizing nitrogen (N) fertilizer management in wheat is essential for maximizing yield and minimizing environmental impact.
• Precision agriculture relies on accurate, real-time crop status monitoring, including nitrogen nutrition.
• Developing effective remote sensing techniques is key to improving wheat nitrogen management.

Purpose of the Study:

• To evaluate the efficacy of different sensors in estimating winter wheat nitrogen status.
• To identify optimal vegetation indices (VIs) and analytical methods for accurate nitrogen indicator estimation.
• To compare the performance of various regression techniques for predicting wheat nitrogen concentration and nutrition index.

Main Methods:

• Four different sensors were employed to collect spectral data from winter wheat.
• Gaussian process regression (GPR) combined with sequential backward feature removal (SBBR) was used to select optimal VIs.
• Parametric regression (PR), multivariable linear regression (MLR), and GPR were utilized to estimate wheat leaf N concentration (LNC), plant N concentration (PNC), and nutrition index (NNI).

Main Results:

• Optical fluorescence sensors demonstrated superior accuracy in estimating wheat N status under low-canopy coverage.
• The Dualex Nitrogen Balance Index (NBI) proved to be the most effective leaf-level indicator for early growth stages.
• Multiplex indices excelled as canopy-level indicators during early growth, while ASD VIs were more accurate at late growth stages.
• GPR with SBBR outperformed MLR and PR in estimating LNC, PNC, and NNI across different growth stages.

Conclusions:

• Sensor selection and analytical methods significantly impact the accuracy of wheat nitrogen status estimation.
• The combination of specific VIs and advanced regression techniques like GPR with SBBR offers a robust approach for precision nitrogen management in wheat.
• This research provides valuable insights for developing improved remote sensing tools for real-time monitoring of wheat nutritional status.