摘要
当前主流的分割算法一般采用的是选取阈值分割法、最大类间方差(OTSU)法或者是区域生长法,而经试验证明,采用选取恰当的阈值分割法能有效地把文字与背景进行分离。理论分析和大量的实验结果均表明,采用合适的单阈值分割方法提取的简牍文字较清晰,文字细节部分保留的较完整,笔画间的粘连现象较少,达到了令人满意的结果,有利于后期简牍文字修复工作和考古研究的顺利进行。
Maximum Between-Class Variance and area segmentation methods are used separately to bamboo slips.Through a large of experiment,a suitable segment method is found for bamboo slips words distilled,which could separate the words from the background;Both the theory analysis and the experimental results show that the bamboo slips words is much more clearly and more words can be known with the method of single threshold values segmentation.The pleasant result is helpful for the next study and archaeology research.
出处
《科技通报》
北大核心
2012年第6期128-129,132,共3页
Bulletin of Science and Technology
基金
国家"十五"科技攻关项目-文物保护关键技术(2004BA810B)
关键词
图像分割
阈值分割
OTSU分割
区域生长
image segmentation
threshold value segmentation
maximum between-class variance（OTSU）
area segmentation
