The aim of this work was to verify usefulness of image analyses software to determine genotype differences on quantitative linseed traits (Linum usitatissimum L.). Also the goal was to propose new quantitative characters of seeds suitable for determination infra-specific differences. The study was realized on seeds of 6 linseed genotypes from two harvested years – 2010 (5 genotypes), 2012 (5 genotypes). The 7 quantitative traits measured on whole seeds and 3 quantitative traits measured on longitudinal seed section were chosen. Seed samples were randomly selected from genetically and physically purified seeds. The experimental data were obtained by the software for image analysis with module for automatic measurement, from detailed digital images. The methodology for measuring parameters in micropyle region of flax seeds was proposed. According to results of descriptive statistics was found out that the smallest seeds had genotype Oural and the biggest seeds had genotypes Flanders and Recital. The three newly proposed characters from micropyle region of seed had shown higher variability than the traits measured on the whole seeds. Coefficient of variation had mostly values more than 10 %. The One-Way ANOVA confirmed statistically significant differences (p < 0.05) between all analysed samples in chosen quantitative traits. The results of the post hoc means comparison revealed that genotype Recital received the highest values of means and significantly differentiated from means of genotype Oural in the traits measured on whole seeds in both harvested years. Also there was found out that the Pearson correlation coefficient was mostly around r = 0.50 (p < 0.001), except two cases among traits - Area - Feret ratio and Diameter – Feret ratio. The correlations between parameters of micropyle seed region were found very weak (r < 0.50; p < 0.001).
