中国古重彩画矿物颜料高光谱定量分析研究

The quantitative retrieval of mixing mineral pigments on Chinese historical heavy-color paintings is of great significance to the conservation and restoration of cultural relics. Hyper-spectral remote sensing is attractive in this field as a non-invasive imaging technique because it is fast and hence capable of imaging large areas of an object giving both spatial and spectral information. Nowadays, hyper-spectral remote sensing has been mainly utilized as a qualitative analysis tool in the discrimination of pure pigments on Chinese historical heavy-color paintings. However, the spectral mixing mechanism of mineral pigments on Chinese historical heavy-color paintings is not clear yet, and which spectral unmixing algorithm is suitable for the quantitative retrieval of mineral pigments needs to be further studied. To address this issue, spectral experiment of mixing mineral pigments on Chinese historical heavy-color paintings will be performed under precisely control of both the mineral compositions and the environmental conditions. Based on this experimental data, the spectral mixing mechanism of mineral pigments on Chinese historical heavy-color paintings will be investigated. Giving full consideration of the rich species of mineral pigments and the influence of the paper background, technologies for retrieval of the components and fractions of mineral pigments will be explored, and then the quantitative retrieval model for mineral pigments on Chinese historical heavy-color paintings will be developed. This proposal will lay the foundations for the promotion of the application capability of hyper-spectral remote sensing on the conservation and restoration of cultural relics in China, due to enabling the accurate extraction of mineral pigment information from Chinese historical heavy-color paintings.

中国古重彩画中的矿物颜料往往由多种矿物成分混合而成，对矿物颜料进行成分解译和定量还原是我国文物保护和修复领域的重要难题。高光谱遥感能够无损提取古画矿物颜料信息，对于该问题的解决有很大潜力。目前，中国古重彩画矿物颜料的高光谱遥感分析以纯净矿物颜料的定性识别为主，在定量分析方面缺少相关研究。因此，本项目拟开展中国古重彩画矿物颜料高光谱定量分析研究，内容包括：设计模拟实验精确控制混合矿物颜料成分含量并模拟古画老化过程，获取混合矿物颜料光谱数据；深入研究中国古重彩画矿物颜料成分光谱混合模型，发展中国古重彩画矿物颜料光谱解混算法；在此基础上构建中国古重彩画矿物颜料定量分析模型，并通过真实中国古重彩画成像光谱数据处理验证模型精度。本研究拟解决中国古重彩画修复中面临的重要技术难题，将为我国文物保护和修复提供新的技术手段和信息支持。

中国古重彩画中大部分成分是混合调配的矿物颜料，运用高光谱遥感技术分析混合矿物颜料的成分和含量仍是一个难题。矿物颜料成分的光谱混合模型是高光谱混合矿物颜料成分分析的基础，但中国古重彩画矿物颜料光谱混合模型尚不明确。因此，探索一种适合于中国古重彩画矿物颜料的混合矿物颜料的成分鉴定与定量分析方法成为文物保护与修复领域的一个迫切需要。本项目针对文物保护与修复的需求，设计模拟混合矿物颜料，深入研究中国古重彩画矿物颜料成分光谱混合模型和定量反演算法。主要成果包括：①在全波段矿物颜料混合分析中，使用全约束最小二乘，广义双线性模型等多种线性和非线性模型进行处理，结果表明矿物颜料光谱混合总体上非线性特征较强，并且包络线去除处理能够在一定程度上减少非线性，提高解混精度；②采用比值导数解混算法进行单波段光谱解混，发现在某些波段矿物颜料的光谱混合符合线性混合模型，具有较强的线性特征，可以用线性解混得到较高精度；③开发了中国古重彩画矿物颜料光谱解混算法，提出了log and CR (LCR)模型，并建立了一种基于混合反射率重建的综合精度评价体系，为光谱解混模型的进一步研究奠定了坚实的基础；④对唐卡主色矿物颜料进行体系性的光谱特征分析，总结了不同色系矿物颜料可见光、近红外、短波红外谱段光谱特征；⑤通过对古代纺织品文物和青花瓷的光谱特征分析，发现短波红外波段范围光谱适用于纺织品纤维类型的鉴别，可见光-近红外波段范围内不同青花料光谱特征有显著差别，为高光谱技术推广应用于纺织品、瓷器等类型文物研究提供了技术支持。通过总结和整理相关成果，已发表SCI收录5篇，北大核心刊物论文1篇、国际会议论文2篇；申请发明专利1项；培养博士生1名、硕士生6名。该成果解决了中国古重彩画修复中面临的重要技术难题，可为我国文物保护和修复提供新的技术手段和信息支持。