Determination of the bruise degree for cherry using Vis-NIR reflection spectroscopy coupled with multivariate...

Area of Science:

• Agricultural Science
• Food Science
• Spectroscopy

Background:

• Cherry quality and marketability are significantly impacted by bruising.
• Objective and rapid methods for bruise detection are crucial for the cherry industry.

Purpose of the Study:

• To determine and classify the bruise degree in cherries using visible and near-infrared (Vis-NIR) reflection spectroscopy.
• To develop a rapid, non-destructive method for assessing cherry bruise severity.

Main Methods:

• Visible and near-infrared (Vis-NIR) reflection spectroscopy (350-2500 nm) was employed.
• Principal Component Analysis (PCA) and Successive Projection Algorithm (SPA) were used for data analysis and optimal wavelength selection.
• Least Square-Support Vector Machine (LS-SVM) models were established for classification.

Main Results:

• A classification accuracy of 93.3% was achieved using LS-SVM with five optimal wavelengths (603, 633, 679, 1083, 1803 nm) selected by SPA.
• Misclassification primarily occurred between normal and slight bruise samples.
• Vis-NIR reflectance showed a correlation with cherry firmness and soluble solids content (SSC).

Conclusions:

• Vis-NIR reflection spectroscopy combined with multivariate analysis provides a rapid and intact method for determining cherry bruise degree.
• This technique supports efficient cherry sorting and enhances postharvest quality control.
• The findings contribute to improving cherry quality, competitiveness, and profitability.