摘要
图像分割是图像处理中最困难的问题之一。针对光学字符识别(optical character recognition,OCR)系统的要求,提出了一种基于边缘检测的多文本页图像分割方法。先检测出由图像中边缘点构成的边缘直线,然后寻找出图像中每个文本页的4条边缘直线所构成的边界矩形,从而可将每个文本页准确分割出来。实验表明,该方法具有较快的速度和很高的正确率,是一种有效的方法。
Segmentation is one of the big problems of digital image processing. According to the need of OCR （optical character recognition） system, an edge detection based on method to segment scanned document images is presented, each of which contain several sheet of papers. After getting the lines of edge pixels, the rectangle of each paper＇s boundary is found out to segment this paper. Experiments show that this approach is fast, accurate and effective .
出处
《计算机工程与设计》
CSCD
北大核心
2007年第8期1861-1863,共3页
Computer Engineering and Design
基金
国家自然科学基金项目(60472002)
关键词
图像分割
边缘检测
边缘直线
直线提取
边界矩形
image segmentation
edge detection
edge line
line extraction
border rectangle
