|Authors: ||T.B. Morgenstern, L.C. Cunha Júnior, N.C.T. Mariani, V. Nardini, G.H.A. Teixeira|
|Keywords: ||LDA, non-destructive method, quality, identification|
The Myrtaceae family has many species and among them stands up the Myrciaria cauliflora (DC) Berg ('Açú') and Myrciaria jabuticaba (Vell) Berg ('Sabará') due to the quality of their fruit which are called jaboticabas.
Although the fruit of both species are quite similar they differ in various qualities attributes such as, fruit size, total acidity and anthocyanin content.
As consumers are more aware in terms of food authenticity the development of techniques that enable rapid, non-destructive classification of fruit in the two species are desirable in the agribusiness chain.
Thus, the aim of this study was to use CIE colour parameters and diffuse reflectance NIR spectroscopy associated with linear discriminant analysis (LDA) to construct models that can differentiate the fruit of these two species.
The principal component analysis (PCA) of CIE colour parameter a* and chromaticity and NIR spectra pre-processing with multiplicative scattering correction (MSC) presented the best clustering results between M. jaboticaba and M. cauliflora fruit.
The a*, chromaticity-LDA model resulted in 70.8% correct classification of fruit and 59% prediction accuracy of the external validation group.
NIR spectra pre-treated with MSC had 95.5% correct classification and 92% prediction accuracy.
Concluding, NIR spectroscopy associated with LDA was more accurate and showed to be a reliable method for the classification and discrimination of intact jaboticaba fruit of both species.
Download Adobe Acrobat Reader (free software to read PDF files)
URL www.actahort.org Hosted by KU Leuven