Image analysis techniques and protocols designed to examine the effect of agricultural harvest practices and textile processing on the white speck phenomenon in dyed cotton yarn.

The principle hypotheses of this research were centered on the expectation that: First, image analysis techniques, based on images obtained with a flat bed scanner could be used to identify and quantify white specks on dyed yam. Second, an operational definition of a white speck utilizing the image analysis techniques and protocols could be developed. Third, within bale variability of white speck could be quantified using dyed yam as the counting media. Fourth, varied textile settings could affect the white speck content of dyed yam. Fifth, agricultural production harvest treatments would affect white specks on dyed yam.

The acquisition of size and shade characteristics with scanner based image analysis techniques allowed a time operational definition to be developed that accurately described specific observable characteristics of white specks on dyed yam. The research also confirmed white specks identified by image analysis were also identifiable by human operators. Once acquired, the image analysis based operational definition was used to facilitate research that examined how various agricultural and textile treatments affect white speck content. Future work utilizing raw fiber as the white speck test media will depend on the amount of variation that exists in commercial bales of cotton. In the sample group, the variation of white speck content within a commercial bale of cotton was found to vary enough to prevent the formulation of predictive models based solely on normal USDA side sampling. In the textile processing phase of the work, textile card production rate was found to be the most significant change factor that had the greatest affect on white speck content. Cotton harvest year and the date of harvest within each year were found to be the most significant of the harvest related factors included in this research.