基于光谱表示和独立成分分析的混合颜料分离方法  

Separation method of mixed pigments based on spectrum expression and independent component analysis

作　　者：王功明[1] 刘志勇[2] 

机构地区：[1]清华大学美术学院信息艺术与设计系,北京100084 [2]中国科学院计算技术研究所,北京100190

出　　处：《计算机应用研究》2015年第2期593-597,共5页Application Research of Computers

基　　金：国家"973"计划资助项目(2012CB725300);中国博士后科学基金资助项目(2013M530035)

摘　　要：提出一种基于光谱表示和独立成分分析的混合颜料分离方法。首先,采用光谱仪获取混合颜料光谱信息,并将其表示为离散信号的形式;然后,对信号进行独立成分分析,得到基本颜料光谱曲线,从而确定基本颜料种类。采用蒙赛尔色卡光谱制作模拟数据,进行三种色卡光谱混合信息的分离实验以及从八种色卡光谱中选择若干种混合后的分离实验,分离出的基本颜料光谱形态与已知的原始颜料光谱形态极其相似,平均相似比为97.64%,最大相似比可以达到99.95%。实验结果表明,该方法能够准确确定基本颜料种类,具有可追溯性,适用于混合颜料分离。This paper proposed a separation method of mixed pigments based on spectrum expression and independent component analysis. First, the spectrometer was used to obtain the spectral information of mixed pigments that was expressed as the discrete signal. Second, the independent component analysis was used to calculate the type of basic pigment. The simulated spectrum data of Munsell color card was used to validate this method. This algorithm could separate the mixed pigment spectrums from three basic pigment spectrums and the mixture from several pigment spectrums within the set of eight basic pigment spectrums successfully. The curve of separated pigment spectrum was very similar to the curve of original pigment spectrum. The average similarity is 97.64%, and the maximum one can reach 99.95%. The experiment results show that this method can determine the composition of mixed pigments accurately, so it is suitable for separation of mixed pigments.

关 键 词：颜料 光谱 独立成分分析 蒙赛尔 相似性 可追溯性 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]