数字碑帖图像的二值化算法研究  

Research on Binarization Algorithm of Digital Rubbing Image

作　　者：马永力[1] 曾瑄[1] 黄志开[1] MA Yongli;ZENG Xuan;HUANG Zhikai(School of Electrical Engineering,Nanchang Institute of Technology,Nanchang Jiangxi 330099)

机构地区：[1]南昌工程学院电气工程学院,江西南昌330099

出　　处：《湖北理工学院学报》2022年第4期18-22,68,共6页Journal of Hubei Polytechnic University

基　　金：江西省教育厅科技项目(项目编号:GJJ161110)。

摘　　要：以中国古代碑帖图像为研究对象,根据光照、边缘、笔画粗细及像素值等位置特征,建立了基于阈值分割和形态学相结合的数字碑帖图像的二值化复原算法。针对碑帖图像的复杂程度,对于光照不均匀的图像,可采用自适应阈值分割;对于背景噪声较多的图像,阈值分割后,进一步采用形态学、连通域和均值滤波相结合的综合算法进行二值化复原。结果表明,综合算法在处理复杂图像及含有大面积噪声的图像时效果理想。Taking the image of ancient Chinese rubbings as the research object,this paper establishes a binary restoration algorithm for digital rubbing images based on the combination of threshold segmentation and morphology according to the characteristics of illumination,edge,stroke thickness and pixel value.In view of the complexity of rubbing images,adaptive threshold segmentation can be used for images with uneven illumination;for images with more background noise,binary restoration is further carried out by a comprehensive algorithm combining morphology,connected domain and mean filtering after threshold segmentation.The experimental results show that the comprehensive algorithm works the best when dealing with images with complex images and images with large area noise.

关 键 词：阈值分割 二值化 连通域 形态学 

分 类 号：TP391.1[自动化与计算机技术—计算机应用技术]