摘要
研究利用计算机辅助甲骨文缀合的有关理论与技术,对于待缀合的碎片自动生成基于"骨版+碎片+特征"三要素的动态疑似目标碎片数据库,在此基础上,基于"时代+字迹+卜辞"三要素研究终级缀合的有关理论和具体实现算法。首先,根据甲骨片图像纹理细节丰富、目标与背景之间差别明显等特点,研究面向甲骨片图像的图像预处理方法;然后,根据甲骨片缀合的实际需求,研究适合于甲骨片缀合的图像轮廓特征、区域特征以及趋势特征提取算法;设计快速高效的甲骨文碎片智能缀合原型系统。
Based on the theory and technique of the computer aided rejoining,when given the bones/tortoise shells that are to be rejoined,we can dynamically generate the database for the possible target fragments,which contains the information about the bone type of the fragments,and the positions where the fragments are originally supposed to be.On top of the database generated,we propose a novel rejoining algorithm based on the age of the bone shell,the style of the character and the explanation of the writing on the shell.First,we pre-process the shell images,with the consideration of the texture of the shells and the difference between the target and the background.Then,we try to extract characteristics of the contour,sector and trend of the shells from the shell images.We design and implement the prototype which can efficiently rejoin the bones/tortoise shells in an intelligent way.
出处
《武汉理工大学学报》
CAS
CSCD
北大核心
2010年第20期194-199,共6页
Journal of Wuhan University of Technology
基金
国家自然科学基金(60973051)
河南省重点科技攻关计划项目(092102210112)
关键词
甲骨文
缀合
计算机
边界匹配
算法
bone/tortoise shells
rejoining
computer
contour matching technology
algorithm