Cucumber powdery mildew detection method based on hyperspectra-terahertz.

Area of Science:

• Agricultural Science
• Spectroscopy
• Information Technology

Background:

• Cucumber powdery mildew significantly impacts crop yield.
• Accurate and early detection of plant diseases is crucial for effective management.
• Traditional disease detection methods can be labor-intensive and subjective.

Purpose of the Study:

• To develop and evaluate a novel method for detecting cucumber powdery mildew using hyperspectral and terahertz imaging.
• To establish effective spectral preprocessing and feature selection techniques for both imaging modalities.
• To build and validate classification models for accurate disease identification.

Main Methods:

• Visible and near-infrared hyperspectral data were preprocessed using wavelet transform, with feature wavelengths selected via stepwise discriminant analysis.
• Terahertz data underwent preprocessing and screening using iterative variable subset optimization - iterative retaining informative variables (IVSO-IRIV).
• Sparse representation classification and linear discriminant models were established for disease identification.

Main Results:

• The hyperspectral model achieved an average recognition rate of 93% for cucumber powdery mildew.
• The terahertz imaging model demonstrated an accuracy of 87.78% for disease identification.
• Analysis of hyperspectral and terahertz feature images provided detailed insights into lesion characteristics.

Conclusions:

• Hyperspectral and terahertz technologies are effective for detecting cucumber powdery mildew.
• The developed methods provide a foundation for applying these technologies to broader crop disease detection.
• This research highlights the potential of advanced imaging techniques in precision agriculture.