摘要
本文提出一种用于抽取字符骨架特征的预处理方案，以摄象机输入的字符图象先被分割出单个字符．然后用ＩＦＡ算法阈化为二值图象，经规格化处理成１６×１２的标准形式后，再用ＣＧＴ算法进行细化，最后得到字符的单象素宽的骨架．
This paper provides a preprocessing method for extraxting character skeleton fea- tures. A character image from the camera is segmented into seperate characters first, then each character is transformed into binary-image by thresholding based on IFA. The 16×12 standard from appears after normalizing. Then CGT method and thinning will be applied. At last a single pixel wide skeleton of character is obtained.
出处
《福州大学学报（自然科学版）》
CAS
CSCD
1990年第3期18-22,共5页
Journal of Fuzhou University(Natural Science Edition)
关键词
字符识别
分割
阈化
细化
预处理
character recognition
segmentation
thresholding
thinning
