Abstract:

Near-infrared diffuse reflectance (NIDR) spectroscopy combined with chemometric techniques was developed
for rapid detection of kudzu starch adulterated with cheaper starches (potato and sweet potato starch). Principal component
regression (PCR) and partial least squares regression (PLSR) were used to build and evaluate calibration models by different
pre-treatment methods and at different wavelength bands . The PLS calibration model based on wavelength bands of 962–
1 389 nm and spectral pre-treatment by the standard normal variate transformation (SNV) showed an excellent prediction
accuracy. The external validation set provided a coefficient of determination (R2), root mean square error of prediction
(RMSEP) and relative prediction deviation (RPD) of 0.994 5, 2.298 7% and 13.56, respectively. The average recovery rate
was 99.89% (n = 9) with a relative standard deviation (RSD) of 2.96%. The results indicate that NIDR spectroscopy can be
used as an efficient way to detect kudzu starch adulterated with cheaper starches.

Key words: near-infrared diffuse reflectance spectroscopy, kudzu starch, adulteration, rapid detection