an innovative approach for authentication of white rice using vis-nir hyperspectral imagingcoupled with multivariate curve resolution and classification techniques

Food fraud or food crime is a significant and expanding global problem [1]. rice, as a staple food for nearly half of the world s population, especially in asian regions, presents a challenge in terms of rapid, reliable, and user-friendly authentication procedures [2]. the quality of rice is closely linked to its geographical origin, which depends on factors such as cultivation soil, climate conditions, and tillage methods [3]. besides health concerns, this challenge is also essential for customers considering the cost. in this study, we propose a visible-near infrared hyperspectral imaging (vis-nir hsi) system combined with chemometric techniques to distinguish among rice samples from three northern provinces of iran. we collected a total of 93 rice samples, with 39 from mazandaran, 38 from gilan, and 16 from golestan provinces. hyperspectral images of the samples in their intact forms were obtained and analyzed using different strategies, including global mean spectra, size reduction of images by binning method and multivariate curve resolution alternating-least squares (mcr-als). among these methods, mcr-als allowed us to extract the pure spatial and spectral profiles as well as pure components of rice grains within the specified wavelength range. the identified spectral profile of the selected component was closely associated with rice pigments, carbohydrates, starch, and protein. by employing principal component analysis (pca) on the resolved spatial profiles (distribution maps), we identified patterns of rice samples. subsequently, we utilized partial least squares-discriminant analysis (pls-da), a supervised classification model, to achieve our objective. by observing the accuracy of 91.2%, 93.3% and 95.15% for each class, the resulting model demonstrated satisfying sensitivity and specificity based on the resolved distribution maps of rice samples, indicating the imaging technique s suitability for analyzing heterogeneous samples. in conclusion, vis-nir hsi combined with chemometric methods proves to be a powerful technique for authenticating the geographical origin of rice. this approach offers a promising solution to the challenges posed by rice authentication, providing efficient and reliable results.