Shape analysis of grains of Indian wheat varieties

Abstract

Shape variation, based on grain morphology, was quantified in 15 Indian wheat varieties by digital image analysis using custom-built software. Gray images of the grains of different varieties were captured in crease-down position using a HP scan jet IICX/T scanner in transparency mode. The software rotates each grain in the captured image for normalization of orientation. These rotated images were used for further analysis. Geometric features such as area, perimeter, compactness, major and minor axis length and their ratios, slenderness and spread were computed on the binary image. Five other shape factors were derived from these basic geometric features. Moment analysis for calculating standard, central, normalized central and invariant moments of grains was done using the gray images of grains. The data for each parameter for every grain in the image as well as mean, standard deviation (S.D.) and standard error (S.E.) for that parameter were stored for further use. The wheat varieties used in this study showed differences in geometric and shape related parameters. It was concluded that Euclidean distance calculated on the basis of these differences could serve as a basis of distinguishing between samples and to their identification.