Assessment Of Tomato Quality Characteristics Using Vis/Nir Hyperspectral Imaging And Chemometrics

The aim of this study was to assess the applicability of VIS/NIR hyperspectral imaging in intelligently detecting postharvest quality of tomatoes. The postharvest quality of tomatoes was evaluated in terms of firmness, soluble solids content (SSC), pH and color (L*a*b*C*h). VNIR and NIR hyperspectral imaging (400-1700 nm) were used to build regression models using Partial Least Square (PLS). The effects of different preprocessing techniques, including moving weighted average smoothing, first- and second-derivative Savitzky Golay (S-G) and standard normal variate (SNV) on prediction performance were also evaluated. NIR spectra-based models performed the best at estimating the quality characteristics in tomatoes. Excellent prediction for SSC, firmness and pH (RPD > 3.0), and good prediction for color (RPD > 2.0) were obtained. Consequently, NIR spectra-based systems could be introduced in the manufacturing process of tomato industries as nondestructive method to monitor quality characteristics of tomatoes.