Accurate discrimination and quantitative prediction of adulterants in matcha using multidimensional fusion information

Matcha powder is a popular food ingredient rich in various beneficial components. Adulteration with low-cost powders seriously infringes consumers’ rights and disrupts market order. Therefore, accurate identification of matcha adulteration is of great necessity. Conventional single detection technologies suffer from insufficient accuracy in identifying adulteration with multi-component mixtures. Accordingly, this study proposes a multi-dimensional information fusion method for the accurate identification and quantification of matcha adulteration. Specifically, near-infrared spectroscopy, laser diffraction particle size analyzer, and an electronic nose were used to obtain data on the physicochemical composition, particle size distribution, and volatile organic compounds of matcha powder, respectively. Subsequently, data fusion was performed to integrate the three-source information, enabling quantitative analysis of matcha adulteration. The results show that the prediction accuracies for matcha adulterated with flavorings, spirulina powder, wheat seedling powder, mulberry leaf powder and mixed adulterants are 90.70%, 93.24%, 94.66%, 93.97% and 93.21%, respectively. These findings demonstrate that the proposed method meets practical application demands and can provide new insights and theoretical support for the intelligent quality monitoring of matcha.