基于特征点最小距离拟合的文档图像倾斜检测  

Document image skew detection based on the least distance fitting of the feature points

作　　者：吴一全[1] 谢静[1] 

机构地区：[1]南京航空航天大学信息科学与技术学院,南京210016

出　　处：《光学技术》2009年第1期152-155,共4页Optical Technique

摘　　要：在文档扫描输入的过程中,文档图像不可避免地会发生倾斜,而布局分析及字符识别对页面倾斜十分敏感,因此倾斜检测和矫正是文档分析预处理中的重要环节。提出了一种基于最小距离拟合的文档图像倾斜检测方法。首先通过对文档图像进行二值化、分块、膨胀运算、连通标记等一系列图像处理操作,以此找到特征点;然后采用最小距离法对这些特征点进行直线拟合,便可获得文档的倾斜角。实验结果表明,该方法能够快速、精确地检测文档的倾斜角度。In process of document scanning input, document images inevitably introduce skew. While the algorithms for layout and character recognition are very sensitive to the page skew, so the skew detection and correction is an important step in the preprocessing for the document analysis. A new method of the document image skew detection is proposed based on the least distance fitting. First a series of image processing operations such as binarization, block partition, expansion and connected components labeling are made to extract the feature points. Then straight-line fitting of the feature points is done by the least distance method. Thus the skew angle of the document image is obtained. The experimental results show that the proposed method can detect the skew angle of the document image rapidly and accurately.

关 键 词：信息光学 倾斜检测 形态学 特征点 直线拟合 最小距离法 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]