一种政府资源档案图像的二值化方法

针对政府资源档案灰度图像，应用线性变换增强图像对比度，利用空域滤波减少噪声对图像的影响，在此基础上，使用改进的大津率法计算图像的整体阈值T2，使用Bernsen算法计算图像的阈值曲面T1，进而定义笔划断裂阈值T3、邻域平均阈值T4、全局和局部相结合的阈值T5；然后综合利用T1-T5五个阈值，对灰度图像进行二值化。实验表明，这种基于全局和局部阈值相结合的方法对政府资源档案具有广泛适应性，对伪影和断笔划处理的效果令人满意。

An Image Binarization Method of Government Document Resource

For the gray-level image of government document resource,a linear transform is employed to enhance the image contrast.A spatial filter is applied to bate image noise.Based on this preprocessing,the global threshold T2 is calculated by modified Otsu method.The threshold surface T1 is computed by Bernsen algorithm.On the basis of T1 and T2,some thresholds such as the stroke breaking value T3,the average value T4 of local processing window and the union value T5 between global and local are defined.Then the gray-level image is binarized through the combination of these five values.Our experiments show that the proposed method,based on the union of global and local threshold,is adaptive to various documents of government resource.It can achieve our purpose of processing of image ghost and broke rupture.