Feasibility of near-infrared spectroscopy to predict aw and moisture and NaCl contents of fermented pork sausages

The feasibility of near-infrared spectroscopy (NIR) for predicting parameters related to the drying process of fermented sausages (water activity (a(w)), moisture, and NaCl contents) was assessed. A FT-NIR spectrometer and two spectra acquisition setups with contact and remote probes were tested. NIR calibration models were developed using 207 samples scanned between 12,000 and 4000 cm(-1) (833-2500 nm) in reflectance mode. Partial least squares (PLS) regression was used to process spectra and develop calibrations. Predictive models for moisture, a(w) and NaCl yielded 0.997, 0.988 and 0.974 determination coefficients in prediction (R-p(2)) respectively with the on-contact probe method, and 0.998, 0.985 and 0.974 respectively with the remote measurement head method. Furthermore 0.675, 0.006 and 0.117 root mean square errors of prediction (RMSEP) were achieved with the contact probe method, and 0.622, 0.007 and 0.116 with the remote measurement head method. The results confirmed that NIR spectroscopy is an useful technique for predicting moisture and NaCl and suggest it could also be useful to predict a(w) on the surface of fermented sausages. Both setups are appropriate for further on-line applications for monitoring drying processes in a non-destructive way with non-significant differences in the predictive accuracy. (C) 2010 Elsevier Ltd. All rights reserved.