Abstract: Due to the heavy workload and time-consuming and material-consuming of traditional manual color measurement and matching, color spinning technology came into being. Color spinning technology is a spinning technology that blends several colored fibers in a specific proportion to produce fashionable colors. The fabric and finished product made by using color spinning do not need to be dyed again and is considered a green ecological short process technology. However, due to the inability to freely control color during the spinning stage, the actual production and application of color spinning are greatly limited. To address this issue, a four primary color ternary coupling superposition full-color gamut grid-based color mixing model was first constructed, which can perform color phase control, brightness control, and chromaticity control within the full-color gamut range.
On this basis, combined with the characteristics of the Kubelka-Munk dual constant theoretical model, 84 grid point mixed sample formulas were selected from the constructed grid-based color mixing model, of which 54 mixed samples were used as measured samples. With a three-channel CNC rotor spinning machine as the platform, four primary colors of cyan (C), magenta (M), yellow (Y), and white (W) were used as raw materials. Based on the constructed full-color domain grid-based color mixing model and color mixing chromatography, we prepared actual spinning samples. Then, we measured the color values of 54 measured samples, and used the least squares method to calculate the K and S values of each primary color fiber, in order to achieve the prediction of full gamut color or primary color fiber mixing ratio. We also selected the remaining 30 mixed samples as prediction samples to verify the ability of the traditional Kubelka-Munk dual constant theory model to predict the color or primary color fiber mixing ratio. From the comparison between the predicted reflectance of the mixed samples and the actual reflectance, it is found that the predicted reflectance of some mixed samples is significantly lower than the actual reflectance. In response to the problem of insufficient prediction accuracy of traditional Kubelka-Munk double constant theory, the article proposes to reconstruct the Kubelka-Munk double constant theory model for color prediction, and then partially replace the part of the traditional method where the obviously mixed color yarn has a lower reflectivity than the actual reflectivity with the interpolation method. The results show that compared with the traditional Kubelka-Munk double constant theory, the average color difference of the reflectance predicted by the new method has been reduced from 1.48 to 1.04, and the color difference of all mixed samples can be controlled within 2.0. We use the least squares method to predict the monochromatic fiber blending ratio of ten mixed samples, and then substitute it into the Kubelka-Munk dual constant theoretical model to calculate the predicted reflectance. According to the CMC2:1 color difference formula, the color difference between the predicted reflectance and the actual reflectance of the ten mixed samples was obtained, with the minimum color difference being 0.18, the maximum being 0.91, and the average value being 0.45. As the mixing ratio changes, the color difference of the mixed samples fluctuates up and down within its average range, and the color difference is small. The prediction effect of the blending ratio is good. This prediction method has better prediction accuracy than the traditional Kubelka-Munk double constant theory. The constructed four primary color grid mixing model and Kubelka-Munk double constant theory model can be applied to predict the color mixing and mixing ratio of multi primary color fibers.

Key words: colored spun yarn, Kubelka-Munk dual constant theoretical, color prediction, rotor spinning, reflectivity

摘要： 为了能在色纺纱的纺纱阶段即时调控纱线颜色，减少混色成本，缩短工艺流程，结合三通道数控转杯纺纱的特点构建了全色域网格化混色模型，该模型可在纺纱过程中进行全色域范围内的色相调控、明度调控和彩度调控。为了解决色纺纱的测配色问题，得到与之相匹配的测配色系统，根据来样快速进行计算机测配色，节约成本，结合传统Kubelka-Munk双常数理论模型的特点，从构建的全色域网格化混色模型中选取混合样来进行颜色预测。从传统Kubelka-Munk双常数理论模型颜色预测的结果发现，部分混合样的预测反射率明显低于实际的反射率，针对这个问题，重新构建了新的Kubelka-Munk双常数理论模型来进行颜色预测得到新的预测反射率，并用插值替换的方法，把传统Kubelka-Munk双常数理论模型预测结果中明显低于实际反射率的部分用新的预测反射率替换，得到最终的混合样预测反射率。结果表明：与传统的Kubelka-Munk双常数理论模型预测混合样颜色结果相比，新的Kubelka-Munk双常数理论模型预测颜色并用插值替换法替换后得到的最终的混合样的颜色，色差平均值从1.48降低到1.04，且所有混合样的色差均能控制在2.0以内。该预测方法较传统Kubelka-Munk双常数理论模型具有更好的预测精度，所构建的全色域网格化混色模型和新的Kubelka-Munk双常数理论模型可应用于多基色纤维混色色彩和混合比的预测。

关键词: 色纺纱, Kubelka-Munk双常数理论, 颜色预测, 转杯纺纱, 反射率