Nondestructive Measurement of Fresh Tomato Lycopene Content and Other Physicochemical Characteristics Using Visible−NIR Spectroscopy

Abstract

Measurement of fresh tomato fruit overall quality, and particularly lycopene content, is challenging in the context of high-volume production. An experiment was conducted to simultaneously measure various quality parameters of tomato in a nondestructive manner using vis-NIR reflectance spectroscopy and chemometrics. The sampling set included different cultivars that are obtainable from both retailers' shelves and two greenhouse producers. Results indicate that lycopene content was accurately predicted [r(2) = 0.98; root mean square error of cross-validation (RMSECV) = 3.15 mg/kg], along with color variables such as Hunter a (r(2) = 0.98), L, and b (r(2) = 0.92). Tomato color index (TCI) was better predicted (r(2) = 0.96) than the a/ b ratio (r(2) = 0.89). Firmness prediction, with an r(2) of 0.75, is comparable to what is reported in the literature for other fruits and may have a practical interest. Prediction of internal quality such as pH, soluble solids, titratable acidity, and electrical conductivity was less accurate, partly due to a low variability of these parameters among samples. Predictions were robust with regard to cultivars, except for pink variety tomato. The 400-1000 nm range gave results almost as accurate as the 400-1500 nm range.