Verification of Commercial Near-Infrared Spectroscopy Measurement and Fresh Weight Diversity Modeling in Brix% for Small Tomato Fruits with Various Cultivars and Growth Conditions

Abstract

The use of non-destructive commercial near-infrared (NIR) spectroscopy to estimate Brix% was verified using all samples of cherry tomato ‘TY Chika’, currant tomato ‘Microbeads’, and the M&S or market-purchased and supplemental local source tomatoes. Additionally, the relationship between fresh weight and Brix% of all samples was examined. These tomatoes had a diversity of cultivars, growing methods, harvest timing, and production locations and varied widely from 4.0% to 14.2% for Brix% and 1.25 g to 95.84 g for fresh weight. Regardless of the diversity of all samples, it was revealed that the refractometer-based Brix% (y) was practically estimated from the NIR-derived Brix% value (x) using a relationship of y = x (RMSE = 0.747 Brix%) after only a one-time calibration for the NIR spectrometer offset. An inverse relationship between fresh weight and Brix% could be modeled using a hyperbolic curve fit, and the model showed an R2 of 0.809 except for ‘Microbeads’. The Brix% of ‘TY Chika’ was highest on average (9.5%) and had a large difference from 6.2 to 14.2% among the samples. Data distribution of cherry tomato groups such as ‘TY Chika’ and M&S cherry tomatoes was closer, indicating a roughly linear correlation between fresh weight and Brix%.