Convenient detection of white pepper adulteration by portable NIRS and spectral imaging with chemometrics

Abstract

As an important condiment, the quality of pepper attracts much attentions. Due to the high illegitimate interests, pepper adulteration has become more and more serious. Non-destructive and low-cost adulteration detection is urgently needed in on-site scenario. As required, a portable NIR spectrometer was attempted to establish the grade recognition models by LDA, SVM, PLS-DA and SIMCA. Subsequently, quantitative modeling was carried out on piperine, and the optimal model obtained the R2c of 0.979, R2p of 0.973, and RPD of 6.05, which achieved an excellent quantitative performance. Moreover, to further visualize the adulteration, samples with different adulterant concentrations were scanned by hyperspectral imaging. The MCR-ALS was used to reduce the dimensionality of multivariate data and execute iterative calculation. Finally, the pixel concentrations in the spectral image were reconstructed to obtain the visual distribution of adulterant. Overall results demonstrate that a complete non-destructive detecting solution was provided for the analysis of adulteration of white pepper.