摘要
光学字符识别(OCR)时,输出的文本行顺序需与实际的顺序相符。文章在字符Blob分析的基础上,对两个字符Blob外接矩形的相对位置关系进行划分,确定了各位置关系下同一文本行的判断方法,据此对排序后的字符Blob进行文本行初次生成和文本行二次合并,实现了任意方向文本行的生成。实际测试结果验证了所述方法生成任意方向文本行的有效性。
The output text line order from OCR(optical character recognition)process should be consistent with the actual order.On the basis of analyzing character Blob,this paper divides the relative position relationship of the bounding rectangle of the two characters Blob to determine the judgment method of the same text line under each position relationship.Then text lines are generated for the first time with the sorted character Blobs and merged for the second time,which realizes the arbitrary direction text line generation.The actual test results verify the effectiveness of the method.
作者
王海丰
Wang Haifeng(Nanjing Bilin Intelligent Identification Technology Co.,LTD.,Nanjing,Jiangsu 210000,China)
出处
《计算机时代》
2022年第3期11-13,18,共4页
Computer Era
关键词
光学字符识别
BLOB分析
外接矩形
任意方向
文本行
optical character recognition
Blob analysis
bounding rectangle
arbitrary direction
text line