Determination of soluble solid content in market tomatoes using near-infrared spectroscopy

Tomatoes are widely consumed worldwide, and the soluble solid content (SSC) is one of the most important quality parameters for the commercialization of fresh tomatoes, mainly in the salad group. In this regard, partial least square models for intact tomatoes SSC were developed using a portable F-750 Vis-near-infrared (NIR) with Zeiss MMS1-NIR spectrometer in an interactance geometry. Thus, tomatoes from five regions (states of Goiás, Bahia, Santa Catarina, Minas Gerais, and São Paulo) were collected weekly from November 2018 to November 2019, with a total sample number of 2.085, divided into three populations, two for calibration and one for prediction. The best partial least squares regression prediction model was obtained using the Vis-NIR spectral region of 840–1050 nm with Orthogonal Signal Correction (OSC) pre-treatment applied. The calibration population standard deviation (SD) was 0.52%, and for the prediction population, the SD was 0.56%. Low root mean square error cross-calibration of 0.32%, and root mean square error prediction of 0.32% were achieved. The models were able to discriminate low from high values and vice versa.