Ta et al. (2025) Research on Water and Fertilizer Diagnosis of Maize Using Visible–Near-Infrared Hyperspectral Technology

⚠️ Warning: This summary was generated from the abstract only, as the full text was not available.

Identification

• Journal: Agriculture
• Year: 2025
• Date: 2025-12-30
• Authors: Na Ta, Yanliang Li, Xiaofang Yu, Julin Gao, Daling Ma, Jian Qiang Chen, Xu Dou
• DOI: 10.3390/agriculture16010084

Research Groups

Not provided in the text.

Short Summary

This study developed and evaluated hyperspectral estimation methods for maize agricultural traits (relative chlorophyll content, leaf water content, leaf nitrogen content) under varying water and nitrogen regimes, finding that Random Forest models achieved high accuracy (R² up to 0.95) for trait prediction.

Objective

• To explore and evaluate spectral estimation methods for agricultural traits (relative chlorophyll content, leaf water content, leaf nitrogen content) in maize leaves under water-saving and fertilizer-reduction strategies.

Study Configuration

• Spatial Scale: Individual maize leaves/plants within an experimental plot.
• Temporal Scale: V12, R1, and R3 growth stages of maize.

Methodology and Data

• Models used: Support Vector Regression (SVR), Random Forest (RF). Spectral transformations: First-derivative, Second-derivative.
• Data sources: Hyperspectral data (collected from leaves), direct measurements of agricultural traits (relative chlorophyll content (SPAD values), leaf water content, leaf nitrogen content). Factorial experiment data (different nitrogen application rates (N0–N4) and irrigation levels (W1–W4)).

Main Results

• Reducing nitrogen by 10% (N3) had no significant effect on physiological indicators.
• Reducing irrigation by 10% (W3) led to significant differences in physiological indicators.
• First- and second-derivative transformations of spectral data enhanced the correlation with agricultural traits.
• Random Forest (RF) models outperformed Support Vector Regression (SVR) in predicting agricultural traits.
• RF model estimation accuracies (R²) were 0.92 for relative chlorophyll content (SPAD), 0.94 for leaf water content, and 0.95 for leaf nitrogen content.

Contributions

• Provides a robust methodology for non-destructive, rapid estimation of key maize agricultural traits using hyperspectral data.
• Demonstrates the superior performance of Random Forest models for this application compared to SVR.
• Offers technical support for real-time growth monitoring and precise water and nutrient management in maize cultivation, particularly under resource-saving strategies.

Funding

Not provided in the text.

Citation

@article{Ta2025Research,
  author = {Ta, Na and Li, Yanliang and Yu, Xiaofang and Gao, Julin and Ma, Daling and Chen, Jian Qiang and Dou, Xu},
  title = {Research on Water and Fertilizer Diagnosis of Maize Using Visible–Near-Infrared Hyperspectral Technology},
  journal = {Agriculture},
  year = {2025},
  doi = {10.3390/agriculture16010084},
  url = {https://doi.org/10.3390/agriculture16010084}
}