摘要
在文本识别预处理中 ,书写线的去除是一重要工作。采用标准Hough变换的主要缺点在于它需要的存储空间量大和计算时间长。提出了改进Hough变换 ,通过空间复用减少了空间要求和重复运算 ,加快了计算速度 ,并结合实际 ,通过限定角度搜索范围和提取特征点 。
The removing of a text line in a document is an important preprocessing job in document recognition. The main disadvantage of standard Hough transform is the massive storage space demands and long time computation. To overcome this shortage, we have developed an advanced method to do the job by diminishing the space demands and reducing the time spending. Some examples in the post envelope have been proposed in this article with the use of this method, By restricting angle and adopting feature pointers, we can detect and remove the line in envelope with improved Hough transform to satisfy the real use.
关键词
图像处理
HOUGH变换
文本处理
书写线去除
image processing
Hough transform
text line extraction
handwritten document
