基于光谱吸收特征分析的彩绘文物颜料识别研究

高光谱成像技术作为一种无损高效的检测方法,对于彩绘文物颜料的鉴别具有重要意义。由于全波段参与光谱相似度计算会造成数据的冗余,没有充分利用光谱的细微特征。为此,本文首先对文物颜料光谱吸收特征进行参量化分析,并通过改进的光谱吸收特征拟合算法与标准光谱进行匹配识别,从而得到识别结果。实验以获取的一幅波长为400~1000nm的古代壁画高光谱影像为例,通过光谱吸收特征分析识别出壁画的颜料主要成分有朱砂、赭粉、石绿和石青,4种颜料的光谱吸收特征拟合度分别是:0.95、0.77、0.92、0.81。实验结果表明:对光谱吸收特征分析可以帮助识别彩绘文物的颜料信息,该方法可为以后文物修复提供参考。

Pigment Identification of Colored Relics Based on Improved Spectral Absorption Feature Fitting Algorithm

As a non-destructive and efficient detection method,hyper-spectral imaging technology has a great significance to the identification of pigment on colored relics.The results show that the spectral similarity calculation of the full band will cause the data redundancy,and do not make full use of the subtle features of the spectrum.In this paper,the spectral absorption characteristics of the pigment were analyzed,and the spectral feature fitting algorithm was used to match the standard spectrum,and the results were obtained.The experimental study took a hyper-spectral image of an ancient mural as an example which wavelength range is 400-1000 nm,and distinguished the pigments by spectral absorption analysis.The results show that the major pigments are cinnabar,ochre powder,malachite and azurite,the spectral fitting characteristics are 0.95,0.77,0.92,0.81 respectively.The analysis of spectral absorption features can help to identify the color information of colored relics.The method can provide reference for cultural relic restoration.