半透明配色建库仿真算法

Simulation Algorithm of Semi-transparent Color Match for Establishing Database

K-M理论指出K/S值与涂料的浓度存在线性关系,但实际中这种线性关系不明显。随着涂料浓度增加,K/S值逐渐趋于一个稳定值,故大部分配色软件都采用对数曲线拟合方式建立数据库。而涂料在半透明状态下,K/S值的线性性更差,采用上述建库方式已经满足不了配色精度要求。本文在半透明配色中采集多个样点后,首先对其进行无效点去除的预处理,然后对剩余有效点进行三次样条插值,并对低浓度点进行反射率修正使其更符合K-M理论假设的线性关系。实验结果表明,通过上述建库方式能显著提高配色的精度。

The K-M theory points that there is a linear relationship between the K/S values and the dye concentration. In practice this linear relationship is not obvious, so most of the software is logarithmic curve fitting way to establish a database. In semi- transparent situation of dye, the linear relationship of K/S is worse, the above mentioned method has failed to meet color accuracy requirements. In this paper, we collected multiple semi-transparent samples and removed invalid points, then did cubic spline in- terpolation for the remaining valid points, and corrected reflectivity in low concentration points to make it more consistent with the linear relationship hypothesis of K-M theory. Experimental results show that the above-described method can significantly improve the accuracy of color match.