波段分区的数码喷墨印花机光谱特性化模型

Spectral reflectance piecewise partition model for characterizing digital inkjet printer

为提高数码喷墨印花技术颜色复制质量,研究了数码喷墨印花机的光谱特性化模型,提出基于光谱反射率的波段分区的数码喷墨印花机特性化模型。该模型针对31维光谱反射率数据与喷墨印花设备驱动值间的非线性关系,按照反射率波段进行分区,在不同波段分区中使用高次多项式拟合喷墨印花设备输出驱动值与颜色光谱数据的转换,通过转换矩阵实现数码喷墨印花设备的光谱特性化,将该光谱特性化模型应用于蚕丝织物输出实验。结果表明:90%的蚕丝织物喷墨印花色样的测量结果与光谱特性化估算结果的均方根误差都小于0. 001 2,其中最大误差为0. 001 9,平均误差值为0. 000 8; 90%的喷墨印花色样与估算结果的色差小于1. 0 NBS,最大色差为5. 709 9 NBS,平均色差为0. 570 0 NBS。

In order to improve the quality of color reproduction of digital inkjet printing,this paper studied the spectral characteristics of the printer which is the key technology for the textile inkjet printing,and presented a spectral reflectance piecewise partition model (SRPPM) for characterizing digital textile inkjet printing.Based on the nonlinear relationship between the 31 dimensional spectral reflectance and the digital driving data of the ink-jet printer,the model partitioned according to the reflectance band,used the nonlinear high order polynomial to fit the characteristic conversion coefficient between the output driving data and the color spectral reflectance in different bands,and applied the spectral characteristics of the digital inkjet printer by the obtained conversion matrix.Experimental results with skill textile show that the 90% spectral errors of testing color samples are less than 0.001 2,the maximum spectral error is 0.001 9,and the average spectral error is 0.000 8.While the 90% color differences of testing color samples are less than 1.0 NBS,the maximum value is 5.709 9 NBS and the average value is 0.570 0 NBS.