基于BP神经网络的全色域彩色纱颜色预测模型构建

Construction of a color prediction model for full color domaincolor yarn based on BP neural network

针对传统测配色理论的局限性,利用神经网络良好的非线性映射能力构建一种全色域彩色纺纱的颜色预测模型。以灰、青、品红、黄作为四基色进行全色域混色模型的构建,并选取该全色域模型中的66个网格点进行纺纱,制备混色纱及织物,利用分光光度计测出66种织物的光谱反射率曲线,建立基于颜色预测混纺比和基于混纺比预测颜色值的神经网络,并根据预测值与实际值之间的色差值和均方误差评估这两种神经网络的准确性。结果表明:基于BP神经网络的颜色预测模型可以实现色纱颜色值与四基色颜色比例之间的非线性映射,测试样本的平均色差在2,后续将使用遗传算法或粒子群优化算法对其进一步优化,提高全色域彩色纱颜色预测模型的准确性。

A color prediction model for full color gamut color spinning yarn is constructed by using the good nonlinear mapping ability of neural network to address the limitations of traditional colorimetric theory.A full color gamut model is constructed with gray,cyan,magenta and yellow as the four base colors,and 66 grid points in the full color gamut model are selected for spinning to prepare blended yarns and fabrics,and the spectral reflectance curves of 66 fabrics are measured by spectrophotometer,neural networks based on color prediction of blending ratio and blending ratio prediction of color values were established,and the accuracy of these two neural networks was evaluated according to the color difference value and mean squared error between the predicted and actual values.The results show that the color prediction model based on BP neural network can realize the nonlinear mapping between the color yarn color value and the proportion of the four base colors,and the average color difference of the test samples is around two.The subsequent plan is to using genetic algorithm or particle swarm optimization algorithm to further improve the accuracy of color prediction model for full color gamut colored yarn.