Location: Cotton Structure and Quality ResearchTitle: A statistical analysis of the intra-sample distribution and variation of cotton color and its application) Author
Submitted to: National Cotton Council Beltwide Cotton Conference
Publication Type: Proceedings
Publication Acceptance Date: 1/31/2013
Publication Date: 5/24/2013
Citation: Cai, Y., Cui, X., Rodgers III, J.E., Martin, V., Watson, M. 2013. A statistical analysis of the intra-sample distribution and variation of cotton color and its application. Proceedings of 2013 National Cotton Council Beltwide Cotton Conference. p. 614-617. Interpretive Summary: As a natural product, Cotton color has a distribution and variation within the same sample, which may indicate the variation of fiber quality that influences the quality of its textile product. Currently the colorimeter method for measuring cotton color provides the overall color grade; but it does not include information about the within-sample color distribution and variation. We conducted an investigation using color image analysis and statistical analysis to study the distribution and variation of cotton color. We present theoretical and experimental work of the study, and the statistical aspects of the distribution and variation of cotton color and the applications. This research is aimed to develop a better measurement for cotton quality.
Technical Abstract: A cotton’s color has a distribution and variation within the same sample (intra-sample variation. Scanned color images were used to compute samples’ Rd and +b values. The results were well associated to HVI measured Rd, +b values. Color grades were determined by using the computed Rd and +b values consequently. For a cotton, the inter-sample color variation is small, but the intra-sample color variation is much higher. The intra-sample color grade has a distribution that spans over a large range of color grades on the color chart. The intra-sample color grade histogram indicates that cottons with the same overall color grade may have different intra-sample Rd and +b distributions and variations. Therefore at least one proper parameter or a combination of parameters characterizing the intra-sample color variation is helpful for better evaluating color properties. The intra-sample color distribution can be described by using the bivariate normal distribution. The intra-sample distributions’ locations, shapes, orientations, and sizes can be quantitatively analyzed, consequently be evaluated with regard to their impacts on cotton color quality.