Discrimination of cultivars and determination of luteolin content of Chrysanthemum morifolium Ramat. using multispectral imaging system

Abstract

Chrysanthemum morifolium Ramat. (Chr) is a notable medicinal and edible crop that stimulates health owing to its radical-scavenging and antioxidant properties due to its main active flavonoid luteolin. The performance of discriminating between Chr cultivars and determining luteolin content in Chr using multispectral imaging (MSI) was investigated. A combination of MSI with principal component analysis (PCA) and least squares-support vector machines (LS-SVM) was applied to classify Chr cultivars. PCA derived from the spectral and morphological features data of the samples explained 99.61% for summing up the first three principal components and the LS-SVM model achieved 98% discrimination accuracy in the prediction set. Additionally, partial least squares (PLS) and LS-SVM models were obtained to predict performance for luteolin content determination, with R_p of 0.949 and 0.965, and RMSEP of 0.387 and 0.314 mg g⁻¹, respectively. All results demonstrated that the combination of MSI system with chemometrics methods could be a rapid and non-destructive method to discriminate between Chr cultivars and determine luteolin content in Chr.