A Spectroscopic Technique for Water Content Determination in Potato

Food grading has always been a research topic because of the variations among food products. Many subjective assessment methods with poor repeatability and tedious procedures are still widely used. Spectroscopy has been introduced for rapid determination of compositional, functional, and sensory analysis of food products. In this study, a spectroscopic technique was developed to determine water content in potato. The study demonstrated high accuracy in water content prediction. Partial least squares regression was used to establish calibration models for predicting the water content in potato samples. Significant wavelength ranges of 700-900, 1000-1100, and 1250-1600 nm and 850-900, 1100-1200, 1400-1500 nm were selected to establish prediction models for estimating the water content in potato samples without skin and with skin respectively. The RMSE was found to be lower for validation set in both the experimental test sets using the selected wavelength segments, thus giving an indication of higher reliability on the calibration models that used selected wavelength segments.