A green and efficient method for detecting nicosulfuron residues in field maize using hyperspectral imaging and deep...

Area of Science:

• Agricultural Science
• Biotechnology
• Data Science

Background:

• Nicosulfuron herbicide residue detection in maize is crucial for effective crop management and safety.
• Current detection methods are often slow, expensive, and labor-intensive, hindering timely interventions.

Purpose of the Study:

• To develop and validate a novel deep learning model (HerbiResNet) for rapid and accurate detection of nicosulfuron herbicide residues in maize.
• To assess the model's performance against traditional methods and explore correlations between spectral data and herbicide presence.

Main Methods:

• Development of the HerbiResNet model utilizing spectral data from maize leaves across various varieties and herbicide concentrations.
• Analysis of residue levels categorized as low, medium, and high.
• Comparison of HerbiResNet performance with Support Vector Regression (SVR), Partial Least Squares Regression (PLSR), Multi-layer Perceptron (MLP), and AlexNet models.

Main Results:

• HerbiResNet achieved a coefficient of determination (R²) of 0.88 for residue prediction and 0.87 accuracy for residue level classification.
• The model significantly outperformed conventional regression and classical neural network models.
• Identified strong correlations between specific spectral bands (550 nm, 680 nm, 750 nm, 1000 nm) and herbicide presence/physiological changes.

Conclusions:

• The HerbiResNet model provides a highly accurate and efficient method for detecting nicosulfuron herbicide residues in maize.
• Spectral technology combined with deep learning shows significant promise for advancing precision agriculture and sustainable farming.
• This approach lays a foundation for broader applications of spectral sensing in agricultural monitoring.