色纺纱的计算机配色研究进展

Research progress of computer color matching for colored spun yarn

针对色纺纱配色原理复杂,难以对其进行颜色预测和配色的问题,归纳分析了几种可用于色纺纱计算机配色的预测模型在应用方面的特点。结果发现:Stearns-Noechel模型、Kubelka-Munk理论、BP神经模型的计算精度较理想,Friele模型精度较低;Stearns-Noechel模型和Friele模型需求解未知参数且参数对预测精度影响较大,Kubelka-Munk理论与理想条件仍有差距且计算烦琐,BP神经网络需要大量训练样本加强泛化能力。最后指出:色纺纱的配色应在继续提高传统配色模型精度的同时,改进求解配色模型中的未知参数,寻求新型配色模型;同时应注重企业个性化的参数设定,以及通过原液着色纤维的标准化,简化配色中的复杂计算,提高色纺纱的计算机配色技术水平。

Aiming at the complicated color matching principle of colored spun yarn and difficult prediction on matched colors,the application characteristics of several prediction models for color matching of colored spun yarn were summarized and analyzed.The calculation precision of Stearns-Noechel model,Kubelka-Munk theory and BP neural model are ideal,and the precision of Friele model is lower.The Stearns-Noechel model and the Friele model,however,need to solve unknown parameters which have a great influence on the prediction accuracy.The Kubelka-Munk theory still has a gap with the ideal conditions and the calculation is cumbersome.The BP neural network requires a large number of training samples to enhance the generalization ability.Finally,it is pointed out that the color matching of color spinning yarns should improve the accuracy of the conventional model,and also improve the method for solving the unknown parameters in the color matching model and seek a new color matching model.Meanwhile,the personalization of the company′s parameter settings should be paid more attention and the coloring fibers of the undiluted liquid should be standardized.The complicated calculation in color matching is simplified,and the computer color matching technology of the color spinning yarn is improved.