Dynamic Nondestructive Detection Models of Apple Quality in Critical Harvest Period Based on Near-Infrared...

Area of Science:

• Agricultural Science
• Food Science
• Spectroscopy

Background:

• Apple quality is often improved by bagging during growth.
• Accurate in-tree nondestructive testing models are vital for apple valuation.
• Internal quality assessment requires reliable methods for selecting raw apple materials.

Purpose of the Study:

• To establish an in situ nondestructive testing model for in-tree apples.
• To compare the effects of low and high maturity on soluble solids content (SSC) detection.
• To optimize SSC detection accuracy using feature selection algorithms.

Main Methods:

• Acquired low- and high-maturity apples using a handheld internal quality assessment tester.
• Employed four feature selection algorithms, including ant colony optimization (ACO), to analyze diffuse reflectance spectra.
• Utilized Partial Least Squares-Linear Discriminant Analysis (PLS-LDA) for fruit maturity prediction based on color data.

Main Results:

• Diffuse reflectance spectra of high-maturity apples showed better correlation with SSC.
• The combination of high-maturity apple spectra and ACO achieved optimal SSC prediction (Rp=0.88, RMSEP=0.5678 °Brix, RPD=2.466).
• PLS-LDA achieved high accuracy in predicting fruit maturity (99.03% for low-maturity, 99.35% for high-maturity).

Conclusions:

• In-tree apple in situ detection shows significant potential for improving quality modeling.
• Diffuse reflectance spectroscopy coupled with ACO is effective for SSC prediction in apples.
• Accurate fruit maturity prediction is achievable using spectral and color data.