[Study on application of Gaussian fitting algorithm to building model of spectral analysis]

Area of Science:

• Spectroscopy
• Chemometrics
• Agricultural Science

Context:

• Spectral analysis is crucial for determining plant physiological parameters.
• Accurate estimation of chlorophyll content in corn leaves is vital for crop management.
• Traditional spectral analysis methods can be complex and data-intensive.

Purpose:

• To introduce and evaluate a Gaussian fitting algorithm for spectral data analysis.
• To simplify spectral information by fitting Gaussian peaks and extracting key parameters.
• To develop and validate a model for estimating corn leaf chlorophyll content using this algorithm.

Summary:

• The Gaussian fitting algorithm decomposes spectral data into Gaussian peaks, extracting parameters like peak height and position.
• This method reduces spectral data dimensionality, converting 1551 absorbance points to 9 Gaussian parameters.
• Models combining Gaussian fitting with Partial Least Squares (PLS) and Principal Component Regression (PCR) demonstrated high correlation coefficients (0.960-0.962) and low relative standard deviations (0.048-0.0485) for chlorophyll estimation.

Impact:

• The Gaussian fitting algorithm significantly improves the accuracy and simplicity of spectral analysis models.
• It offers a practical and feasible approach for quantitative analysis, outperforming conventional methods.
• This technique enhances model interpretability and efficiency in estimating plant biochemical components.