Plant leaf chlorophyll content retrieval based on a field imaging spectroscopy system

Area of Science:

• Agricultural remote sensing
• Plant physiology
• Spectroscopy

Background:

• Accurate estimation of crop physiological parameters like chlorophyll content is crucial for precision agriculture.
• Field imaging spectrometer systems (FISS) offer potential for non-destructive, high-resolution spectral data acquisition.
• Evaluating FISS performance for quantitative spectral analysis in agriculture is necessary.

Purpose of the Study:

• To verify the performance of FISS for quantitative spectral analysis.
• To estimate chlorophyll content in soybean leaves using FISS data.
• To determine optimal quantitative spectral analysis methods for FISS data.

Main Methods:

• Utilized a field imaging spectrometer system (FISS) with 344 bands (380-870 nm) to collect spectral data from soybean leaves.
• Employed multiple linear regression (MLR), partial least squares (PLS) regression, and support vector machine (SVM) regression for chlorophyll content retrieval.
• Compared spectral reflectance and derivative reflectance, and analyzed the impact of regression methods on retrieval accuracy.

Main Results:

• Derivative reflectance proved more sensitive for chlorophyll content estimation than spectral reflectance, reducing root mean squared error (RMSE) by 3.3%-35.6%.
• Regression methods had a minimal impact on retrieval accuracy compared to spectral features.
• A multivariate linear model achieved the lowest RMSE (0.201 mg/g), a >30% reduction compared to non-imaging ASD spectrometer data, indicating high accuracy.

Conclusions:

• FISS provides high-quality spectral and spatial information, suitable for quantitative spectral analysis in agriculture.
• Derivative reflectance and multivariate linear models are effective for retrieving chlorophyll content from FISS data.
• FISS demonstrates significant potential for widespread application in the agricultural sector for crop monitoring.