Study on image analysis application for identification Quinoa seeds (Chenopodium quinoa Willd) geographical provenance

Abstract

In quality control procedures of biological materials, the geographical provenance is an important information to know. Image analysis is an accessible method that can provide this data. In this work, the geographical provenance of 25 varieties of quinoa seeds, cultivated in Europe and South America was identified from 38 averaged measurements as geometric dimensions, morphological features, fractal dimensions and color features. Principal component analysis (PCA) and hierarchical cluster analysis carried out with all of data allowed to distinguish the five varieties cultivated in Europe. They were located in one sector of the tridimensional PCA scatter plot, where 75% of the variability are expressed, and in one cluster of the UPGMA dendrogram with a cophenetic correlation coefficient of 0.7. © 2009 Elsevier Ltd. All rights reserved.