成像光谱技术的古画隐藏信息提取
发布时间：2018-10-17 10:08
【摘要】：目的古画隐藏信息是包含在画作中的一种微弱信息,对于文物鉴定与修复具有非常重要的意义,如何快速无损准确提取古画中微弱的隐藏信息是目前研究和应用的一大技术难题,本文利用高光谱成像技术快速无损提取古画隐藏信息。方法对古画进行高光谱扫描成像,得到短波红外波段(1 000 2 500 nm)的高光谱立方体数据,联合多种光谱匹配技术分析古画中人物头冠颜料种类;利用主成分分析技术,提取古画头冠周围涂抹痕迹。结果通过多种光谱匹配算法对颜料进行分析,确定了颜料种类,并对匹配算法进行了对比,结果表明在颜料混合的情况下,光谱匹配算法仍能够有效的识别出混合颜料的端元种类,光谱信息散度与光谱角的匹配算法SID_SA匹配精度可以达到0.096,相比其他匹配算法,效果最好;利用主成分分析技术分层剥离主要背景信息后,头冠周围位于画作底层的涂抹信息得到了加强,可以有效地区分背景颜料信息与涂抹信息,提取隐藏信息。结论高光谱分析技术对古画隐藏信息的提取有很好地识别效果,能够准确提取绘画颜料种类,发现绘画过程中的涂抹等痕迹,特别适用于混合颜料的种类识别与画作底层信息提取,发掘画作隐藏信息,为文物保护修复与鉴定提供支持。
[Abstract]:Objective the hidden information of ancient paintings is a kind of weak information contained in paintings, which is of great significance for the identification and restoration of cultural relics. How to extract the weak hidden information from ancient paintings quickly and accurately is a big technical problem in current research and application. This paper uses hyperspectral imaging technology to extract the hidden information of ancient paintings quickly and without loss. Methods the hyperspectral cube data of short-wave infrared band (1 000 2 500 nm) were obtained by hyperspectral scanning imaging of ancient paintings, and the types of crowns pigments in ancient paintings were analyzed by means of multiple spectral matching techniques, and principal component analysis (PCA) technique was used. Extract the smear marks around the crown of the ancient painting. Results the pigments were analyzed by various spectral matching algorithms, and the types of pigments were determined, and the matching algorithms were compared. The results showed that when the pigments were mixed, Spectral matching algorithm can still effectively identify the end-member types of mixed pigments. The matching accuracy of spectral information divergence and spectral angle matching algorithm SID_SA can reach 0.096, compared with other matching algorithms, the effect is the best. Using principal component analysis (PCA) technology to delaminate the main background information, the smear information located in the bottom layer of the painting around the crown is strengthened, which can effectively distinguish the background pigment information from the smear information, and extract the hidden information. Conclusion the extraction of hidden information of ancient paintings by high spectral analysis has a good recognition effect, and it can accurately extract the types of painting pigments, and find the traces of painting, such as smear, etc. It is especially suitable for the category recognition of mixed pigments and the extraction of the underlying information of the painting, and to excavate the hidden information of the paintings, and to provide support for the preservation, restoration and identification of cultural relics.
【作者单位】： 中国科学院遥感与数字地球研究所;广州星博科仪有限公司;
【基金】：北京市科技委员会项目(Z151100003615004)~~
【分类号】：K879.4;O433;TP391.41
本文编号：2276317
[Abstract]:Objective the hidden information of ancient paintings is a kind of weak information contained in paintings, which is of great significance for the identification and restoration of cultural relics. How to extract the weak hidden information from ancient paintings quickly and accurately is a big technical problem in current research and application. This paper uses hyperspectral imaging technology to extract the hidden information of ancient paintings quickly and without loss. Methods the hyperspectral cube data of short-wave infrared band (1 000 2 500 nm) were obtained by hyperspectral scanning imaging of ancient paintings, and the types of crowns pigments in ancient paintings were analyzed by means of multiple spectral matching techniques, and principal component analysis (PCA) technique was used. Extract the smear marks around the crown of the ancient painting. Results the pigments were analyzed by various spectral matching algorithms, and the types of pigments were determined, and the matching algorithms were compared. The results showed that when the pigments were mixed, Spectral matching algorithm can still effectively identify the end-member types of mixed pigments. The matching accuracy of spectral information divergence and spectral angle matching algorithm SID_SA can reach 0.096, compared with other matching algorithms, the effect is the best. Using principal component analysis (PCA) technology to delaminate the main background information, the smear information located in the bottom layer of the painting around the crown is strengthened, which can effectively distinguish the background pigment information from the smear information, and extract the hidden information. Conclusion the extraction of hidden information of ancient paintings by high spectral analysis has a good recognition effect, and it can accurately extract the types of painting pigments, and find the traces of painting, such as smear, etc. It is especially suitable for the category recognition of mixed pigments and the extraction of the underlying information of the painting, and to excavate the hidden information of the paintings, and to provide support for the preservation, restoration and identification of cultural relics.
【作者单位】： 中国科学院遥感与数字地球研究所;广州星博科仪有限公司;
【基金】：北京市科技委员会项目(Z151100003615004)~~
【分类号】：K879.4;O433;TP391.41
【相似文献】
相关期刊论文 前2条
1 杨小儒;张良;王希军;刘玉;;一种高分辨率短波红外宽温度范围被动消热差光学系统[J];激光与光电子学进展;2012年01期
2 李新;张国伟;寻丽娜;谢萍;洪津;郑小兵;;短波红外平场光谱仪的波长定标[J];光学学报;2008年05期
相关硕士学位论文 前1条
1 张飞翔;短波红外激光非视域三维成像系统及性能研究[D];哈尔滨工业大学;2016年
，本文编号：2276317
本文链接：https://www.wllwen.com/shekelunwen/kgx/2276317.html
