| For textile products,the first thing that catches your eyes is the color.At the same time,the color is also the most appealing.Therefore,the color spun yarn with soft color and unique layered style won the favor of consumers.On the other hand,in order to comply with the mainstream of today's era,the environmental protection of color spinning determines its bright future.However,there is still a problem of difficult color matching in the production of colored spun yarns.Most enterprises still adopt the color matching of artificial experience.This method is not only affected by the psychological factors of color matching personnel,but also has the problems of complicated color matching procedures,low efficiency and high cost.In response to these problems,based on three different color models and color spinning database,this paper developed a computer color matching system suitable for color spinning.First of all,for the hemp ash yarn,we carried out on the BP neural network model and built.a three-color color model.Three stimulus value of the standard sample was used as input layer for training,and the network training result was good.After that,five additional samples were selected for verification,and all the fitting errors were less than 1,which showed that the model could be applied to the color matching of hemp gray yarn in the color measurement system of the color spinning machine.Secondly,on the basis of Friele model,based on the method of determining unknown parameters in the previous study,100 sets of standard samples with different color mixing ratios were spun to determine an optimal fixed parameter.Then randomly selected 14 samples for verification,the average fitting color is greater than 1,the color results are not ideal.Therefore,the author improves on the method of determining unknown parameters.The same 14 groups of samples are selected for calculation.The parameter Q is iterated in the interval of[0 1],and the optimal parameters are selected.The average color difference of the fitted formulas is 0.399.This shows that the improved color matching accuracy has been greatly improved,while eliminating the need for a large number of previous experiments,but there are still more than 1 color samples.Therefore,based on the visual characteristics of the human eye,the reflectance at 31 different wavelengths is given different weight coefficients,and 36 groups of samples(including the previous 14 groups of samples)are selected for color matching.Get the average fitting color difference is only 0.2626,and all the fitting color difference is within 1,indicating that the color matching results are very good.Thirdly,aiming at the Stearns-Noechel model,a new evaluation criterion is proposed,that is,the relative deviation of the formula can directly reflect the closeness between the fitted formula and the real formula.After that,based on the previous studies,we try to change the selection criteria of unknown parameters,and choose the corresponding parameter M from the minimum color difference to the minimum reflectance.The result of the calculation is that all the fitted chromatic aberration is less than 1,and the relative deviation of the formula is reduced,which shows that the improvement result is better.Finally,this subject chooses Datacolor company's SF600 color measuring instrument as the color system development hardware support.Based on the above three improved algorithms,a smart color matching module is set up,and the experimental data in this topic is built using SQL Sever to build a color spinning database module.Using MATLAB to design a visual color program interface and connect with SQL Sever,the color spinning computer color matching system was developed.Using this topic developed computer color matching system software,by the reflectance spectrum curve of the color spun yarn,the proportion of the colored fiber formulation of each component of the color mixing yarn is predicted.This can improve color matching efficiency and reduce color matching costs. |
PDF Full Text Request