一种有效的文本图像二值化方法

An Effedtive Binarization Algorithm For Document Image

针对一般文本图像二值化方法--全局阈值法和局部阈值法所存在的不足,提出了一种整体与局部相结合的二值化方法,该方法根据文本图像的特点,自适应调整局部二值化的窗口大小,能有效避免断笔,伪影等现象,同时保持较快的处理速度。

This paper proposes a new method for binarization with global and local information combined, aiming at the weakness of the single global and local threshold. Because of the characteristics of document images, the new method extracts the window size for local binarization adaptively, which not only avoids disconnection and ghost but also remains a fast speed.