基于Euclidean距离的纸张z向截面中墨层图像分割

目的通过图像分割方法得到纸张z向截面中印刷墨层区域的图像,利于更好地分析研究油墨在纸张中的分布。方法利用物理切割方法和超景深显微镜拍摄,得到高分辨率的纸张z向截面彩色图像,将该图像从RGB颜色空间转换到HIS空间,采用欧式距离作为相似性度量,实现油墨层的分割。结果在HIS空间对油墨层图像进行分割时,分割阈值选取为3倍样本标准差最大值时,分割图像精确;当分割阈值取4倍或者2倍、1倍时,出现了过度分割或者欠分割。结论分割得到的油墨层图像有利于定量计算油墨的渗透深度及渗透分布,为进一步定量研究油墨与纸张的相互作用奠定了一定的基础。

Segmentation of the Ink Layer in z-direction Cross-section Views of Offset Paper by Using Euclidean Distance and Digital Microscope

The aim of this study was to analyze the distribution of printing ink in the paper more precisely by segmenting the ink layer image from the z direction cross section view of paper. The physical cutting method was used to cut the paper samples and a digital microscope with extra depth of field was used to capture the high resolution views of z-direction cross-section of offset samples printed with cyan ink. After transforming the z-direction cross-section view from RGB color model to HIS color model, the ink layer image was segmented using Euclidean distance as the similarity measurement. By using the HIS image and the threshold of three times of T, the optimum segmentation result of ink layer was got and was valuable for the further quantitative research on the penetration of ink into the paper. When the threshold was set four, two or one time of T, oversegmentation or insufficient segmentation occurred. The penetration depth of ink in the paper can be accurately calculated from the segmentation result of ink layer which can be applied to further researches on the Interaction between ink and paper.