高光谱成像技术在彩绘文物分析中的研究综述

彩绘文物是文化遗产研究的重要内容之一。目前,许多的化学、光谱以及数字成像等分析技术应用于彩绘文物研究中,其中,高光谱成像技术集光谱分析与成像技术为一体,具有无损、快速成像以及"图谱合一"的特点。其技术特点使得高光谱成像技术在非接触、无样本的条件下对彩绘文物进行无损研究,既可以获得彩绘文物的整体形貌特征,还可以深入分析彩绘文物的光谱特征,是高光谱成像技术相比于其他彩绘文物研究方法的独特优势。利用高光谱成像技术研究彩绘文物分为数据采集、数据分析以及数据应用三步,其中数据分析与数据应用是研究的主要内容。通过对高光谱成像技术在彩绘文物中的相关研究成果进行总结归纳,其数据处理方法主要包括高光谱数据降维、光谱特征参量化、光谱解混合以及分类方法四个方面,并分别描述了四类处理方法的主要功能、常用方法和已有案例。从具体应用方向上,可归纳为视觉增强、隐含信息挖掘、保护监测和颜料分析四类,具体描述了四类应用方向所涵盖的内容以及所解决的问题。最后对相关研究中存在的挑战和发展前景进行了总结和展望。

Review of Hyperspectral Imaging Technology in Analyzing Painted Artifacts

The painted artifacts is an important part of the cultural heritage.At present,chemical,spectral analysis and digital imaging technologies have been applied in the research of painted artifacts.It is true that hyperspectral imaging technology set spectral analysis and imaging technology have the advantages of being nondestructive,fast imaging and "image-spectrum merged".Because of these technical characteristics,hyperspectral imaging technology functions under the condition of non-contact and without sample for the non-destructive researching of painted artifacts.Meanwhile,the unique advantage of hyperspectral imaging technology compared to other analysis technique of painted artifacts is that the overall morphology characteristics and the in-depth spectral e characteristics of painted artifacts both can be obtained from hyperspectral images.There are three steps in the research of painted artifacts with hyperspectral imaging technology including data acquisition,data analysis and the application of data.Among them,data analysis and application are the main content of the study.Through summarizing the related researches of hyperspectral imaging technology in painted artifacts,the data processing methods are analyzed from four aspects:hyperspectral dimension reduction,spectral feature parameters,hyperspectral unmixing and classification method.The main functions,common methods and case studies of four data processing methods are described.The application of related researches can be summed up in four directions:vision enhancement,revealing hidden information,production monitoring,and pigment identification,which is described from the content and solved problems.In the end,the challenges of existing related research are summarized,and and carries on the prospects.