摘要： 提出了一种新的自适应二值化方法。基于解决全局二值化方法分割图像的整体轮廓和细节不能兼顾的问题,首先,提取图像细节部 分区域作为子像块,并对每一块小区域进行全局二值化;然后,再对整个图像进行全局二值化。由于这些区域非常小区域二值化对全局二值,化的结果的影响可以忽略不计。另外,提出了利用边缘提取后的图像作为提取区域的依据。实践证明,这种二值化方法能够取得非常好的效果,并且具有相当强的场景适应能力。

关键词: 自适应二值化, 边缘提取, 连通域, 区域生长

Abstract: It is offered that a new adaptive method to segment images in this article. In order to solve the problem that custom binary images don't have good performance on both whole figure and details.On the first step,get the small areas occupying the image details,and segment these small areas in original image,and then,change the whole image into binary image.Because the areas are very small,the method will not do great harm to the whole binary image.What's more,the image after edge-detection is made as the basis of getting these little areas. The experiment proves that this method for both the details and the fundamental figure can get satisfactory result. It can also bring out good performance in similar situations.

中图分类号: 

• TP391.4