Feasibility of fraud detection in milk powder using a handheld near-infrared spectroscopy

Food fraud had become a global food safety concern. Food adulteration is the top of the concern list when talking about food fraud. For example, the melamine incident obtained significant attention from worldwide due to its misuse as a food adulterant. It is a nitrogen-rich substance always added purposely to food such as milk powder to increase the apparent protein content and the profit without using expensive materials. It may lead to serious illness or death if the melamine content exceeds the permitted amount. Thus, the demanding safety and quality of the milk powder increased. This research will study effective and rapid fraud detection in milk powder by using near-infrared spectroscopy (NIR). Melamine was added to milk powder with different concentrations and detected it by near-infrared spectroscopy. The total of 130 spectral data from 10 adulterated samples and 30 unadulterated samples with 3 different lighting condition. The results data will be interpreted by using Principal Component Analysis (PCA) and Logistics Regression (LR) to discriminate between adulterated and unadulterated samples. From the results, detection accuracy of 100 % can be achieved on training and independent test set after spectra pre-processing. The results show the feasibility of NIR with the combination of PCA and LR as it can use as a fast screening tool for fraud detection in milk powder.