用于染色的神经网络计算机配色算法

Study on neural network computer color matching for dyeing

分析了现有染色配色理论的现状,提出了基于神经网络的计算机配色算法,并进行了仿真试验.从仿真过程和结果看,BP网络具有学习能力,对于更新的工艺环境、新的染料系列、新的配方数据可以随时重新训练配色系统,从而实现配色系统的自动调整;还可以改变输入节点和输出节点数目,适应不同采样设备获取的色信息.简化了配色的算法,简化了基础数据的建立,有效减小了配方的色差.

The present theory of the color matching in dyeing industry was analyzed. A new computer color matching method based on neural network was brought forward, and it was used in simulation experiment. Simulation process and the results showed that BP network had learning ability. In order to achieve automatic adjustment of color matching system, it could re-training the color matching system according to the updating environment, the new dye series and new recipe data. It could also change the node numbers of input and output to adapt to the color information of obtained by different sample devices. Thus, the algorithm of color matching and the building of data base were simplified, and the the color difference of formula was effectively reduced.