利用多光谱成像技术, 讨论了一种对书画图像进行扫描和分析的系统修复方法, 避免了传统化学、物理分析方法对书画造成不必要的损害。在计算机的辅助下对书画进行多光谱图像的采集、光谱反射率重建、光谱图像降维以及图像颜色重建等, 实现了书画颜色的高保真再现, 在颜色修复上提出了基于Kubelka-Munk理论的颜料重构方法, 为书画的修复提供技术上的支持。采用该方法不仅可降低修复过程中书画损坏的风险, 而且可提高修复质量。

This paper discusses a systematic repairing method of using the multispectral imaging technology for scanning and analyzing the painting images to avoid unnecessary damage from the conventional chemical and physical analysis. The multi-spectral image acquisition, spectral reflectance reconstruction, spectral image dimensionality reduction and image color reconstruction have been carried with the aid of computers. A high fidelity reproduction of the pigment from the painting and calligraphy is realized. Based on the Kubelka-Munk theory, a pigment reconstruction method is proposed for the color restoration, providing the technical support for repairing the paintings. The proposed method can not only reduce the risk of paintings damage during the painting repairing but also improve the repairing quality, which has far-reaching significance on the heritage of the Arts.