Journal
COMPUTERS AND ELECTRONICS IN AGRICULTURE
Volume 61, Issue 2, Pages 126-135Publisher
ELSEVIER SCI LTD
DOI: 10.1016/j.compag.2007.10.001
Keywords
agro-biodiversity preservation; computer image analysis; flax seed descriptors; flax cultivar clustering; hierarchical clustering; Multivariate analysis; principal component analysis; seed color; seed shape
Ask authors/readers for more resources
We applied computer image analysis to group together flax cultivars (Linum usitatissimum L.) according to their similarity in commercially important dry seed traits. Both the seed shape and seed-color traits were tested on 53 cultivars from world germplasm collections. Four shape traits (Area, Perimeter, MeanChord, and MinFeret) and three color traits (L*, a*, b* calculated from original RGB color channels as CIE color space coordinates) were computer extracted from digital images of 62349 seeds with 1200 seeds per cultivar in average. Cultivar clustering was generated by two independent methods of multivariate analysis. Principal Component Analysis (PCA) was complemented by hierarchical clustering with Unweighted Pair Group Method with Arithmetic Mean (UPGMA). Significant multivariate clustering was obtained both by using non-reduced data set composed of all seven seed traits or by reduced data set made of only three color traits, while calculation with data set of only four seedshape traits did not produce significant cultivar clusters. Based on the results we recommend that current qualitative sensorical seed descriptors routinely used for cultivar characterization may be supplemented by more informative continuous quantitative descriptors obtainable at low cost from dry flax seeds. (c) 2007 Elsevier B.V. All rights reserved.
Authors
I am an author on this paper
Click your name to claim this paper and add it to your profile.
Reviews
Recommended
No Data Available