Research On Computer Color Matching For Textile Dyeing Based On Interpolation By Segments And Iteration Algorithms

General speaking, traditional methods to realize pigments recipe prediction for textile dyeing are spectrophotometric and colorimetric matching which are based on Kubelka-Munk theory. But assumptions in this theory make the methods based on it larger calculation errors. So the recipe obtained is not accurate enough to meet the product demand. In view of the characteristics of color theory and color dyeing mixed, the essence of color matching for textile dyeing is a problem of nonlinear modeling with experimental data. So many researchers try to use nonlinear process modeling effective means-computer color in computer color matching and achieved many results. But generalization and local minimum in traditional neural network make many difficulties for us to solve.In order to overcome the deficiencies of the existing color matching methods, the concept of interpolation by segments and iteration algorithms are introduced to computer color matching processes. And the comparison of calculated and experiment shows that the matching method in this paper is feasible.On the basis of the analysis of the data, the relationship between the concentration of sample from one pigment and the CMY value and the relationship between the concentration of sample from three pigments and the CMY value are researched respectively. According to the law of discovery from the research above, the mathematics model is established based on the affecting coefficient of the assumptions. Then using Sub-function relationship and Interpolation by Segments, the method verifies the affecting coefficient of the assumptions.At last, according to Interpolation by Segments and Iteration Algorithms, the mathematics model resolve the concentration of the sample.Through this research, the mathematics model based on interpolation by segments and iteration algorithms to solve computer color matching problems is more accurate, more effective and more practical than the BP neural network. It can basically meet the color requirements of industrial production.The method proposed in this paper provide new references for textile dyeing color matching and also have certain values for theory study and application research.