摘要
提出了一种对畸变字符图像快速变形校正的算法,他主要是通过对字符主方向的搜索来求取倾斜图像的有关参数。该算法计算形变区域的水平及垂直差分并设计出判别函数,进一步搜索出使判别函数达到最大的主字符方向并进行校正,最后给出了一种图像毛刺的剔除方法。实验表明,该算法充分利用了畸变字符图像的特征,同时使用了较优化的快速搜索策略,取得了较好的实际结果。
In this paper,authors propose a fast algorithm for skew remedy based on searching main direction to acquire the parameters of the skew image. The algorithm computes the horizontal and vertical difference of distorted image,and proposes judge function to find the main direction which makes the judge function achieve maximum, so as to correct the distorted image. Finally some smoothing procedures are used to erase the burrs of characters in the remedied image. Experiments result that the algorithm introduced in this paper thoroughly utilizes the expected results have been derived. the features of skew image and uses the optimal search strategy.
出处
《现代电子技术》
2007年第11期112-114,共3页
Modern Electronics Technique
基金
新疆教育厅基金资助项目(XJEOU2004S35)
关键词
主方向
变形校正
倾斜图像
搜索策略
main direction
skew remedy
distorted image
search strategy