古代瓷釉光学相干层析图像的纹理特征分析研究  

Research on Texture Characteristics Analysis of Ancient Glaze Optical Coherence Tomographic Image

作　　者：严鑫[1] 董俊卿[1] 刘卫东[1,2] 李青会[1] 

机构地区：[1]中国科学院上海光学精密机械研究所科技考古中心,上海201800 [2]武汉东羽光机电科技有限公司,湖北武汉430073

出　　处：《中国激光》2015年第5期241-247,共7页Chinese Journal of Lasers

基　　金：国家973计划(2012CB720906;2012CB720901);青年科学基金项目(51402326);上海市研发平台专项项目(13DZ2295800)

摘　　要：为了对古代瓷釉的结构特征进行量化表征,基于图像的统计直方图以及灰度共生矩阵提取出了7个纹理特征参数,用来对4种不同类型典型素面瓷釉样品的光学相干层析(OCT)图像进行描述。对比分析发现4种瓷釉的7个纹理特征参数差异明显。同时利用未知样品对基于K-邻近分类和7个纹理特征参数的瓷釉自动识别方案进行了可行性验证,得到了未知样品的正确的分类结果。实验结果表明:7个特征参数可以很好地表征瓷釉的釉层结构特征,基于K-邻近分类和纹理特征参数量化分析的瓷釉快速自动识别方案是可行的,且具有广泛的应用前景。In order to quantify the structural characteristics of ancient porcelain glazes, seven texture characteristic parameters, which are used to describe the optical coherence tomographic （OCT） images of four typical plain enamel samples, are extracted based on image histogram and gray level co-occurrence matrix. The differences between the seven texture characteristic parameters of four kinds of glazes are obvious. The feasibility of glaze recognition scheme based on K-neighbor classification and quantitative analysis of texture feature parameters is certified by using an unknown sample. Experimental results show that seven characteristic parameters can characterize the glaze layer structure characteristics commendably. The glaze recognition scheme based on K- neighbor classification and quantitative analysis of texture feature parameters is feasible and promising.

关 键 词：测量 光学相干层析 纹理分析 数字图像分析 瓷釉特征 

分 类 号：TN247[电子电信—物理电子学]