| In color spun,the undyed and colored fibers are blended and carded,then spun into yarn.The fabrics are colorful,varied in style and three-dimensional,meeting people's pursuit of fashion and individuality.And the production process is energy-saving,emission-reducing,green,and environmentally friendly,which is in line with the future development trend of the country.Fiber is a kind of soft,flexible and non-uniform translucent material,which is not easily stabilized,resulting in inaccurate color measurement results.The complexity and diversity of colored fiber blends and the specificity of the fiber's structure result in unsatisfactory prediction results from color prediction models.The complexity and diversity of the colored fiber blends and the specificity of the fiber structure leads to unsatisfactory prediction results of the color prediction model.At present,the color measurement results of fiber samples are inaccurate and the prediction results of the color prediction model Friele model,Stearns-Noechel(S-N)model,Kubelka-Munk(K-M)theory are not satisfactory.As a result,manual color matching is still commonly used,which is subjective,unstable,time-consuming,and poorly reproducible.To solve the problems above,a standardized fiber sample color measurement method is proposed in this paper.Based on the optimized color prediction model,a new algorithm is proposed to improve the accuracy of the color matching results.The research contents and conclusions are as follows:(1)A standard fiber color measurement method is developed by optimizing the fiber sample parameters,to solve the inaccurate color measurement results of fiber samples.The optimized parameters for fiber sample include fiber layer thickness,fiber density,fiber arrangement,and fiber blending uniformity.And the optimal parameter values are determined by analyzing the color change trends of the fiber samples under different parameter settings.Finally,through sample preparation and fiber sample parameter setting,the fiber sample is controlled in a quantitative,moderate thickness,constant density and uniform distribution.The standard fiber color measurement method can ensure the accuracy,stability and repeatability of the color measurement results of fiber sample,and provide an accurate data basis for the study of fiber color mixing rules,which can be applied in computer color measurement and color matching.(2)In order to improve the unsatisfactory prediction results of the colored fiber color prediction model,the accuracy and applicability of the existing color prediction model are compared and analyzed.The mapping relationships between color parameters of colored fibers,mass proportions and color parameters of blended fibers constructed by Friele model,S-N model and K-M theory are explained in principle.Compared with the Friele model and S-N model,the K-M double constant theory can be more adequately explain the intrinsic color connection between colored and blended fibers using three parameters(K?S?R).There is no error accumulation effect in the K-M theory,and the applicability is better.For the three models,the color prediction results of the multi-component colored fiber blending samples are compared and evaluated.The results show that the accuracy of the prediction results by the K-M double constant theory is better than that of S-N model and Friele model.For 87.5%of the multi-component colored fiber blending samples,the CIEDE2000 color difference is less than 2.The K-M double constant theory is used for further research.(3)The absorption coefficient K and the scattering coefficient S of colored fibers in the K-M double constant theory cannot be accurately measured and the solution process is complicated.To solve this problem,the K-M double constant theory is optimized and the solution method of the undetermined parameters K and S is improved.In the optimized model,the reflectance R of the colored fibers can be obtained by color measurement,then the unknown parameter K is characterized by both the unknown parameter S and the known parameter R.In this way,the unknown parameters in the model are halved and the complexity of the solution is reduced.A system of equations between the unknown parameters S and known parameters R of pre-colored fibers and the fiber blends,mass proportion of the fiber blends are constructed to solve the unknown parameters S.The experimental results show that97.5% of the multi-component colored fiber blending samples,the CIEDE2000 color difference is less than 2 in the prediction results of the optimized K-M double constant theory method.(4)To improve the prediction accuracy,a hybrid of least squares and grid search method is proposed to predict the mass proportions of pre-colored fibers in the target sample based on the optimized K-M double constant theory.Since the result of the least squares method is an approximate solution,which satisfies the local optimum but not the overall optimum and the constraint that the sum of the mass proportion of each component is 1.The least squares method lacks the local search capability.For the hybrid method,grid search is used to search and calculate the overall optimum of mass proportion of the target sample based on the approximate solution,improving the accuracy of the prediction results and satisfying the constraint settings.The experimental results show that the hybrid method improves the accuracy of the prediction results of the target sample compared with the least squares method,and the hybrid method improves the color matching efficiency compared with the grid search method.All of the target samples,CIEDE2000 color difference is less than 0.5 in the prediction results of the hybrid method.Based on the above research work,this paper proposes a measurement and matching system for colored fiber blending,which integrates the color measurement,color prediction and data management.The system is equipped with standard color fiber color measurement methods,which can realize fast and accurate color prediction of color fiber blending and target sample mass proportion prediction,and provides a solution for digital storage and management of enterprise product information,showing good application prospects. |
PDF Full Text Request