Utilizing near infrared hyperspectral imaging for quantitatively predicting adulteration in tapioca starch

Fraud creates huge problems for the food industry. One type of fraud is adulteration in order to reduce costs and increase profitability. Fraud occurs in the starch industry, which is difficult or impossible to detect by visual inspection. Therefore this study was to test a possible nondestructive method that could be used to detect the adulterants in tapioca starch by utilizing reflectance near infrared hyperspectral imaging (NIR-HSI) at wavelengths in the range of 935–1720 nm. Pure tapioca starch was adulterated with limestone powder at 0.5% intervals over the range of 0–100% (wt/wt). The samples (n = 201) were divided into a calibration set (n = 140) and a prediction set (n = 61). Chemometrics was investigated and used to establish a calibration model for predicting the concentration of adulterant using partial least squares regression (PLSR). The accuracy of prediction using the model gave excellent results with the correlation coefficient (R) of 0.996 and the root mean square error of prediction (RMSEP) of 2.47%. The model was then used to create the predictive images of pure tapioca starch, adulterated tapioca starch and pure adulterant. It showed different colors based on the concentration of the adulterant. Therefore, NIR-HSI was shown to have potential as a method for rapidly detecting the level of concentration of adulterant in tapioca starch using both the predictive model and visualization.