摘要
传统书画作品中存在有大量模糊不清的印章,对此类印章尚没有很好的手段进行提取和鉴别。为解决这一问题,本工作采用一种用高光谱图像系统采集书画中模糊印章的光谱-图像信息,采用最小噪声分离变换(MNF)处理光谱图像数据,提升了模糊印章的可辨识度。结果表明,采用该方法能够有效的将模糊印章的信息提取出来,有利于印章的鉴别和研究,为书画的文物价值和真伪鉴别的研究提供了科学有效的手段。
Many traditional painting and calligraphy works have blurred and defective seals,but there has been no good way to extract and identify these seals.This paper reports a method of collecting spectrum-image information from blurred seals in calligraphy and painting works using hyperspectral image system.The minimum noise fraction(MNF)rotation was used to process spectral image data,which improved the recognition of blurred seals.The results show that this method can effectively extract the information from the blurred seals,contributing to the research on the value and authenticity of calligraphy and painting cultural relics in a scientific and effective way.
作者
周新光
沈骅
吴来明
ZHOU Xinguang;SHEN Hua;WU Laiming(Key Scientific Research Base of Museum Environment(Shanghai Museum),State Administration for Culture Heritage,Shanghai 200231,China;Conservation Center of Shanghai Museum,Shanghai 200231,China)
出处
《文物保护与考古科学》
北大核心
2020年第1期56-60,共5页
Sciences of Conservation and Archaeology
基金
上海市科学技术委员会科研计划项目资助(17DZ1205100)。
关键词
光谱图像
最小噪声分离
模糊印章
信息提取
书画文物
Spectral image
Minimum noise fraction
Blurred seal
Information extraction
Cultural relics of calligraphy and painting
