Optimized Stearns-Noechel Model to Predict Mixed Color Values of Yarn-Dyed Fabrics

In the textile industry, an innovative approach to achieving a special color is to use only a few colored yarns to gain various fabric color effects. This technology is highly environmentally friendly and can greatly reduce wastewater pollution caused by fabric dyes. In practice, however, matching the color of a woven fabric with the targeted or desired color by interweaving colored yarns experimentally has been difficult. Base on the optical color mixing theory and previous research on pre-colored fiber blending, this paper proposes a new model with an adjustable parameter M to predict the mixed color of yarn-dyed fabric made of dope-dyed polyester filaments. Experiments were carried out on 96 fabric samples that were a single color in the warp direction and different colors in the weft direction. A best-fit algorithm was used to determine that the optimum value of M for the Stearns-Noechel (S-N) color prediction model was 0.021. Differences in the values of the predicted and actual colors as determined with a spectrophotometer were calculated for three models : K/S, log (K/S), and S-N. The predictions of these models showed that the average color differences of the optimized S-N model were lower than those of the other two fixed-formula models. In addition, the optimized S-N model has different prediction accuracy in different range of wavelengths.