Authentication of rapeseed variety based on hyperspectral imaging and chemometrics

The seed authentication is crucial for quality and yield. The traditional detection methods are often destructive, time-consuming and laborious. In this study, authentication of rapeseed variety was proposed by hyperspectral image (HSI) and a partial least squares discriminant analysis (PLS-DA). Hyperspectral imaging of single rapeseed was acquired. Random frog was used to select the important variables, and PLS-DA was used to build a classification model. The validation results based on an independent test set indicated that this model could differeniate the target rapeseed variety from other one. Moreover, to extend the use of this model in practice, the rapeseed samples adulterated with 4 % and 6 % rapeseeds of other varities were prepared to validate this model. The results indicated that this model could also identify the adulteration with other vatities. Subsequently, seed purity was correctly determined by percentage of authentic rapeseeds. In summary, hyperspectral imaging combined with PLS-DA effectively determine the purity of rapeseed. This study provides a reference for rapid seed authentication of other seeds to improve breeding efficiency and optimize germplasm resources.