敦煌壁画中的服饰色彩承载着重要的历史与文化信息，但长期风化与人为干预使其色彩失真严重，亟需科学复原。 本文基于可见光光谱分析，梳理敦煌服饰图像特征与常见颜料类型，分析其劣变路径与光谱变化机制，在精准采样与数据建 模的基础上提出一套可行的色彩复原路径。文中结合莫高窟典型案例，明确色彩复原的多重目标，平衡“材质真实”与“视 觉真实”，并构建多方案并行与可解释性机制，为敦煌服饰色彩的数字再现与文化阐释提供方法支持与实践参照。

The costume colors in Dunhuang murals carry important historical and cultural information, yet long-term weathering and human intervention have caused severe color distortion, making scientific restoration an urgent task. This paper, based on visible spectral analysis, sorts out the visual characteristics of Dunhuang costume images and the commonly used pigments, analyzes their deterioration pathways and corresponding spectral variation mechanisms, and proposes a feasible color restoration framework grounded in precise sampling and data modeling. Taking representative cases from the Mogao Grottoes as examples, the paper clarifies the multiple objectives of color restoration, seeks a balance between “material authenticity” and “visual authenticity”, and establishes a multi-scheme and interpretable mechanism, so as to provide methodological support and practical reference for the digital reproduction and cultural interpretation of Dunhuang costume colors.