运用特征点的矢量空间数据盲水印算法

Feature points based blind watermarking approach for vector data

针对矢量空间数据水印攻击的特点,该文提出了基于特征点的矢量空间数据盲水印算法。在水印嵌入前,应用道格拉斯-普克方法,提取矢量数据特征点;对特征点进行离散傅里叶变换,将水印通过量化的方法嵌入离散傅里叶变换后的幅度系数和相位系数中;水印被多次嵌入,实现了水印的盲提取。试验结果表明,该方法引起的空间数据误差很小,同时对旋转、平移、要素排序、数据格式转换、裁剪、压缩、增删点等攻击具有较好的鲁棒性。

In this paper, a new blind watermarking algorithm for vector data based on feature points was proposed. The characteristics of attacks on spatial data were considered. Douglas-Peucker method was used to get the feature points before embedding watermark. Then DFT transformation was used in the fea- ture points. Watermarks were embedded several times in the coefficients and phase coefficients by DFT transformation. No original data was needed in the watermarks extraction procedure. The experimental re- sults showed that the spatial accuracy of watermarked data was acceptable and the algorithm was robust a- gainst a various of attacks such as rotation, translation, reordering, data format conversion, cropping, compression, vertex insertion and removal.