基于统计特征的二维矢量地图鲁棒水印算法

A Robust Watermarking Algorithm for 2D Vector Maps Based on Statistic Detection

数字矢量地图的实际应用受到数据安全性的限制，数字水印技术是解决该问题的重要辅助手段．在二维矢量地图数据中嵌入水印需要充分考虑地图数据及其应用环境的特殊性．现有算法所暴露的一个共同的缺陷是水印的嵌入会破坏原始数据的形状特征．本文提出了一种基于数据统计特征的鲁棒水印算法，该算法以地图曲线特征点提取为基础，水印的嵌入能够尽量保持曲线本身的形状特征．实验表明该水印方案具有较好的隐蔽性和鲁棒性，能够抵抗地图几何变换、数据简化、插值，以及小幅度的加性噪声．

The application of 2D vector maps faces some security related problems which limit the usage of the map data sets. The technique of watermarking can provide potential solutions for it. Unlike in general multimedia data sets, hiding data in vector maps should take into account some special features of the map data such as the data structure and the application context. A common weakness existing in most previous algorithms is that the watermark embedding procedure often disturbs the shape information of the original map objects （polylines and polygons）. A robust and shapepreserving watermarking scheme based on statistic detection is proposed in this paper. The feature points of the map objects are firstly extracted and the embedding steps will not introduce obvious shape distortions into the cover map. The experimental results show that the proposed scheme has a better performance on invisibility. It can also be demonstrated that the proposed scheme is robust to some attacks such as map simplification, interpolation, rotation, translation and additional noise with low amplitude.