Visual Spectroscopy and Colour Modelling for Nondestructive Quality Evaluation of Apple

Abstract

Potential of visual spectroscopy and colour data obtained using simple handheld colourimeter was investigated for nondestructive prediction of two important internal quality parameters of apple: total soluble solids (TSS, degree Brix) and titrable acidity (TA). Apple colour in terms of Hunter lab L, a, b values and visual spectral (400 to 700 nm) reflectance values of five cultivars were determined. Empirical models were fitted to L, a, b values and partial least squares and multiple linear regression models were fitted to spectral reflection data. The model predictions were validated via full-cross validation technique. Statistical analyses of the results indicate that TA is better predicted than TSS or TA/TSS ratio using colour values. The models based on colour values are better than those using spectral reflectance data. A linear model with colour parameters a, b and a² fitted the experimental TA values with R values of 0.767 and 0.679 for calibration and validation, respectively.