ARS | Publication request: Measure the color distribution of a cotton sample using image analysis

Submitted to: National Cotton Council Beltwide Cotton Conference
Publication Type: Proceedings
Publication Acceptance Date: February 21, 2012
Publication Date: April 16, 2012
Citation: Cui, X., Rodgers III, J.E., Cai, Y., Martin, V., Watson, M. 2012. Measure the color distribution of a cotton sample using image analysis. National Cotton Council Beltwide Cotton Conference. p.1223-1227.

Interpretive Summary: Cotton color is one of the most important factors in cotton trade and textile processing. At present the most commonly used measurement of cotton color only reports the sample’s overall color grade; the color distribution and variation within the sample are not reported. Obtaining the color distribution of a cotton sample enables a more comprehensive evaluation of its color characteristics. In this paper we report our preliminary results on using an image analysis method to study the intra-sample cotton color distributions based on high resolution color images. We examined cotton samples’ intra-sample and inter-sample variations of Rd, +b values, which were used to determine color grade. Our results show that the image analysis method we used effectively obtained the color distributions of the samples. The results also show that though inter-sample color variation of a cotton is very small, the intra-sample variation is significant.

Technical Abstract: The most commonly used measurement of cotton color is by the colorimeter principal that reports the sample’s color grade. However, the color distribution and variation within the sample are not reported. Obtaining color distributions of cotton samples will enable a more comprehensive evaluation of cotton color characteristics. We report our preliminary results on using image analysis as a method to study cotton color distributions. High resolution color images were obtained by using a color scanner. For each image, overall Rd and +b values were computed and compared to those obtained from HVI measurement. Rd and +b values of sub-areas of every image were computed to analyze the color distribution. Intra-sample and inter-sample variations of Rd, +b values and color grades were examined. Results show that though inter-sample color variation of a cotton is very small, the intra-sample variation is significant.