[A path-length correction method on biochemical parameter nondestructive measuring of folium]

Area of Science:

• Agricultural Science
• Spectroscopy
• Biochemistry

Background:

• Visible/Near-Infrared (Vis/NIR) spectroscopy offers rapid, non-destructive analysis of plant biochemical parameters.
• Variations in leaf path-length due to scattering and thickness affect spectral analysis precision.
• Accurate biochemical analysis is crucial for plant health monitoring and agricultural applications.

Purpose of the Study:

• To develop and validate an improved path-length correction method for Vis/NIR spectroscopy.
• To enhance the accuracy of quantitative analysis models for biochemical parameters in plant leaves.
• To address the challenge of spectral variations caused by differing sample path-lengths.

Main Methods:

• Deduction of the Extended Multiplicative Scattering Correction (EMSC) algorithm, incorporating chemical terms and wavelength functions.
• Experiment 1: Assessing path-length effects on spectra using samples with varying thickness and consistent chlorophyll content.
• Experiment 2: Evaluating the EMSC method's efficiency using Partial Least Squares (PLS) regression on samples with diverse thickness and chlorophyll content.

Main Results:

• EMSC preprocessing reduced spectral variations, bringing spectral coefficients close to the spectrometer's repeatability error.
• The root mean squared error of prediction for chlorophyll content improved from 3.9 SPAD to 2.2 SPAD.
• The number of principal components increased from 5 to 12, indicating better model complexity and accuracy.

Conclusions:

• The improved EMSC method effectively eliminates spectral differences caused by leaf path-length variations.
• This preprocessing technique enhances spectral data sensitivity and the precision of calibrated models for biochemical analysis.
• The study demonstrates a significant advancement in non-destructive plant biochemical analysis using Vis/NIR spectroscopy.