Application of Handheld Near Infrared Device for In-Plant Quality Assessment of Tomato Paste Samples

Abstract

The current study aims to evaluate the applicability of a novel handheld NIR scanner and an NIR sensor in transflectance mode for non-destructive and rapid in-situ analysis of important quality parameters in tomato paste samples. The predictive variables included five key quality traits-natural tomato soluble solids (NTSS), titratable acidity (TA), Bostwick consistency, serum viscosity, and the a/b ratio. Reference levels of these parameters were determined using conventional analytical techniques. A total of 224 tomato pastes samples, supplied by a tomato processing industry from 2015 to 2020 and in 2022, were considered for this study. The samples provided a unique range of concentration for individual parameter (NTSS = 25.7-37.4 (0)Brix, TA = 1.01-1.95 %, Bostwick consistency = 1.0-10.4 cm, serum viscosity = 190.0-452.0 cSt and a/b ratio = 1.9-4.4). A transflectance approach was employed to collect NIR spectra using a 0.50 mm pathlength reflector. Partial least squares regression (PLSR) was used to analyze the multivariate data and develop predictive models for the quality traits. Both, the NIR scanner (1350 nm-2500 nm, R(2)pre = 0.83 to 0.98 and RMSEP = 0.03 to 0.63) and the NIR sensor (1100 nm-2550 nm, R(2)pre = 0.85 to 0.98 and RMSEP = 0.05 to 0.43) exhibited comparable performance with good figures of merit (RER >9 and SEP/SECV = 0.8-1.1), emphasizing their suitability for quality assessment of tomato paste samples. Additionally, model transfer from the NIR scanner to the NIR sensor was investigated, and the performance of the resulting multivariate calibration transfer was compared with that of the NIR sensor. Models generated using the NIR sensor performed better than the calibration transfer models, however, the validation performance of the calibration transfer models (R(2)pre = 0.55 to 0.94 and RMSEP = 0.06 to 0.68) suggested their suitability for screening calibration. Overall, the results underscore the reliability of miniaturized NIR systems to streamline quality analysis of tomato paste samples in the tomato processing industry, providing a cost-effective, high-throughput and multicomponent monitoring technique.