Abstract
Predictability of maturity using quality attributes based on Vis-NIR spectra will be beneficial to farmers and consumers alike. Hand-held Vis-NIR spectrometers are a convenient, rapid, non-destructive method that can measure the quality attributes of many fruits and vegetables. The aim of this study is to evaluate the potential of a hand-held Vis-NIR spectrometer to classify the maturity stage and to predict the quality attributes of strawberry such as lightness (L*), chroma colour (C*), hue (H degrees), total soluble solids (TSS), titratable acidity (TA) and total polyphenol content (TPC). Principal component analysis (PCA) was used to distinguish strawberry at different maturities. Partial least squares regression (PLSR) models of internal quality attributes were developed in the spectral region between 550 and 900 nm for a hand-held NIR instrument. Several pretreatment methods were utilized including standard normal variate (SNV), multiplicative scatter correction (MSC), Savitzky-Golay algorithm smoothing and second derivative. Different pretreatment methods had effects on the classification performance of the PCA model. In general, SNV gave better results than the other preprocessing techniques. The coefficient of determination (R-2) of the PLSR (SNV) model was calculated as 0.92, 0.93, 0.92, 0.96, 0.91 and 0.90 for L*, C*, H degrees, TSS, TA and TPC, respectively. Given the importance in assessing strawberry quality at different maturity stages, the use of a hand-held spectrometer, which are usable and rapid, should be considered a non-destructive analysis of strawberry quality.