摘要
本文针对手写体数字的二值化图象提出了一种快速细化算法。本文基于一般的剥离技术,在连通性判别准则,边界点检测、扫描方式等三个方面提出了新的观点和改进措施,使本算法比现有的一些算法快两倍以上。该算法在IBM-PC/XT上用Pascal语言编程实现,具有如下的细化速度,一幅26×45的二值图象(笔划宽度3～5个象素宽)需0.28s的处理时间。文中给出部分实验结果。
In this paper, based on common peeling technique, a fast thinning algorithm for handwritten numeral is presented. In order to raise thinning speed the author make some improvements in connectivity criteria, boundary detecting and scanning modes. Programming in PASCAL language, the algorithm is implemented on IBM-PC/XT with such thinning speed as, a 26×45 bilevel image being processed in 0.28 second. In this paper some experimental results are also presented.
出处
《数据采集与处理》
CSCD
1990年第1期13-17,共5页
Journal of Data Acquisition and Processing
