www.old.international-agrophysics.org / issues


International Agrophysics
publisher:Institute of Agrophysics
Polish Academy of Sciences
Lublin, Poland
ISSN: 0236-8722


vol. 20, nr. 4 (2006)

previous paper back to paper's list next paper
Multivariate analysis of image descriptors of common wheat (Triticum aestivum) and spelt (T. spelta) grain infected by Fusarium culmorum
(get PDF)
E. Suchowilska , M. Wiwart
University of Warmia and Mazury, Department of Plant Breeding and Seed Production, 10-724 Olsztyn, pl. Łódzki 3, Poland

vol. 20 (2006), nr. 4, pp. 345-351
abstract The response of five spring common wheat and five spring spelt cultivars to the infection of their spikes with Fusarium culmorum was examined using the shape and colour analysis of kernel images. The results obtained suggest that there is a significant correlation between the thousand kernel weight (TKW) and the shape descriptors of the kernel image: area, perimeter, length and width. Such a correlation was observed especially for TKW and image area (Pearson's correlation coefficient r ranged from 0. 737 for common wheat to 0. 914 for spelt) as well as for the shape coefficient S5 calculated on the basis of image length and area (r equalled 0.716 and 0.886, respectively). A significant correlation was also observed for TKW and H (hue), S (saturation) and I (intensity) of colour of the kernel image. The results of cluster analysis performed for the values of H, S, I and S5 permitted precise differentiation between kernels obtained from control and infected heads of common wheat and spelt. A reliable evaluation of grain infection by F. culmorum was possible only when the results of both shape and colour analysis were considered.
keywords Fusarium head blight, grain, image analysis, wheat, spelt