1673-159X

CN 51-1686/N

一种改进的文档图像倾斜检测算法

An Improved Skew Detection Algorithm for Document Images

  • 摘要: 在分析现有倾斜检测算法基础上, 提出一种改进的文档图像倾斜检测算法。该算法首先对文档图像进行图像增强、去噪和二值化处理, 然后提取边界线、计算每条边界线的角度, 最后计算加权平均得到文档图像的倾斜角度。通过在计算每条边界线的角度过程中引入LMS算法。改进的算法提高了原算法的精确度、稳定性和适应性。实验表明, 改进的方法有效可行。

     

    Abstract: Based on the analysis of existing algorithms, an improved skew detection algorithm for document is proposed.Firstly, enhancement, denoising and binarization of an image are performed.And then, the extraction of the border line and computation of each borderline's angle are done.Finally, the skew angle is obtained by computing the weighted average.By applying LMS algorithm, the proposed algorithm has better accuracy, stability, and adaptation.The experimental results show that the improved algorithm has better performance.

     

/

返回文章
返回