中文摘要：

针对当前灰度共生矩阵（GLCM）的4个主要特征参数与图像子块纹理复杂度之间没有准确的数学关系和隐蔽图像遮蔽性上不足的问题，提出一种GLCM纹理特征选块的可逆图像水印。首先把原宿主图像分成128×128大小的子块；然后利用均方误差（MSE）给4个纹理特征参数赋予权值，建立特征参数与图像子块复杂度的数学关系；最后计算得出各个子块的复杂度，选择复杂度最大和次大的子块，采用基于预测误差对扩展的可逆算法进行数据隐藏。提出的特征参数与纹理复杂度之间的数学关系，能够准确计算图像子块复杂度，嵌入水印后的自然图像和医学图像平均峰值信噪比（PSNR）值较现有方法分别高出2．65％和0．93％左右。本文算法能准确反映子块内部的纹理复杂度并具有更好的隐蔽性，适用于医学、军事和卫星等领域。

英文摘要：

To solve the problem that there is no clear mathematical relationship between four main characteristic parameters of gray level co-occurrence matrix （GLCM） and texture complexity of image sub- block, a reversible image watermarking algorithm for GLCM texture feature selection is proposed. Firstly, the original host image is divided into 128 × 128 sub blocks. Then the mean square error is used to assign the weight of the four texture feature parameters. The mathematical relationship between the characteristic parameters and the complexity of image sub block is established. Finally,the complexity of each sub block is calculated,and blocks with the maximum and sub maximum texture complexity are select- ed,and the pairwise prediction error is used to hide the data. The mathematical relationship between the characteristic parameters and the texture complexity is proposed,which can accurately calculate the complexity of image sub-block. Compared with existing method, the proposed algorithm improves the average peak signal to noise ratio （PSNR） value of watermarked nature image and medical image respectively by 2.65 % and 0. 93 %. Experimental results show that, compared with previous similar schemes, the algorithm not only can accurately reflect the texture complexity of the sub block, but also has better concealment. The presented method can be applied in many fields, including the areas of medicine, military and satellites.