Prediction of Soluble-Solid Content in Citrus Fruit Using Visible-Near-Infrared Hyperspectral Imaging Based on...

Area of Science:

• Agricultural Engineering
• Food Science
• Spectroscopy

Background:

• Citrus fruit quality is assessed using external and internal factors.
• Non-destructive methods are crucial for efficient post-harvest evaluation.
• Visible and near-infrared (VNIR) hyperspectral imaging offers rapid quality assessment.

Purpose of the Study:

• To evaluate VNIR hyperspectral imaging for predicting citrus soluble solid content (SSC).
• To develop and optimize a predictive model for SSC in citrus fruits.

Main Methods:

• Utilized a VNIR hyperspectral imaging system (400-1000 nm).
• Acquired hyperspectral images from citrus fruits in two orientations.
• Developed a partial least-squares regression (PLSR) model, incorporating competitive adaptive reweighted sampling (CARS) for wavelength selection and outlier removal.

Main Results:

• The CARS-PLSR model achieved a coefficient of determination (R²) of approximately 0.75.
• The model demonstrated a root mean square error (RMSE) of approximately 0.56 °Brix.
• Optimized model performance through spectrum preprocessing, effective wavelength selection, and outlier detection.

Conclusions:

• VNIR hyperspectral imaging is a viable non-destructive technique for predicting citrus SSC.
• The developed CARS-PLSR model shows potential for real-time quality assessment.
• Findings support applications in agricultural and food post-harvest management systems.