Research on SPAD Estimation Model for Spring Wheat Booting Stage Based on Hyperspectral Analysis.

Area of Science:

• Agricultural Science
• Remote Sensing
• Spectroscopy

Background:

• Advancements in agricultural informatization necessitate precise, non-destructive crop monitoring methods.
• Spectral technology offers a powerful tool for assessing crop health and nutrient status.

Purpose of the Study:

• To develop an efficient and accurate method for estimating relative chlorophyll (SPAD) in spring wheat during the booting stage.
• To explore the utility of fractional-order differential transformation and spectral indices for enhanced SPAD detection.

Main Methods:

• Acquisition of hyperspectral reflectance data from spring wheat.
• Application of fractional-order differential calculus to transform spectral data.
• Correlation analysis to identify significant spectral features and indices.
• Development and comparison of estimation models using least-squares support vector machine (LSSSVM) and slime mold algorithm-optimized LSSSVM (SMA-LSSSVM).

Main Results:

• The 0.4 order fractional-order differential spectra showed the highest correlation with SPAD, exceeding original spectra by 9.3%.
• Two-band differential spectral indices demonstrated higher sensitivity to SPAD than single differential spectra, with a maximum correlation of 0.724.
• The SMA-LSSSVM model, utilizing two-band fractional-order differential spectral indices, outperformed other models in SPAD assessment.

Conclusions:

• Fractional-order differential spectral analysis significantly improves the correlation with SPAD.
• Optimized spectral indices derived from fractional calculus are effective for precise chlorophyll estimation.
• The SMA-LSSSVM model provides a robust approach for non-destructive SPAD assessment in spring wheat.