【作者】 孙鸿睿；

【导师】 朱建军；

【作者基本信息】 中南大学 ， 测绘科学与技术， 2013， 博士

【摘要】 随着矢量地图在实际应用中的不断深化,矢量地图的版权保护及数据认证等安全问题日益凸显。数字水印技术作为矢量地图版权保护的重要手段之一得到了广大专家学者、科研机构广泛的关注,本文以矢量地图无损数字水印技术为主题,研究了矢量地图无损水印框架模型、针对如何提高矢量地图无损水印算法的鲁棒性、水印容量等研究了一系列矢量地图无损数字水印算法,本文主要研究成果总结如下：(1)矢量地图无损数字水印技术分析。对矢量地图数据的内容特征、数据组织结构特征进行了深入剖析,在此基础上给出了矢量地图无损水印的通用模型,分析了矢量地图无损数字水印技术的嵌入位置,并根据矢量地图的数据特点,研究了矢量地图数字水印算法评价指标。(2)研究了矢量地图的零水印技术。根据矢量地图的数据特点,给出了矢量地图零水印的框架模型,分析了生成有意义零水印的理论基础。针对传统水印算法抵抗攻击鲁棒性较差的问题,分别从矢量地图的空间特征和属性特征入手,设计了两种零水印方案：一种基于矢量地图特征点和分块的零水印算法,通过提取矢量数据的特征点、对地图进行改进的四叉树(MQUAD)分块处理,设计了一种有效对抗地图几何变换(平移、旋转、缩放等)、数据压缩、顶点删除、增加以及数据裁剪等操作的鲁棒的零水印算法；一种基于矢量地图要素编码的零水印方案,分析了属性作为构造特征信息的可行性,利用要素编码的稳定性构造零水印,从而有效提高了算法在抵抗几何变换(平移、旋转、缩放)、地图简化尤其是地物编辑等方面的鲁棒性。(3)提出了一种改进的矢量地图差值扩张可逆水印算法。差值扩张算法具有算法简单,水印容量较大等优点。在传统差值扩张算法的基础上,对顶点坐标计算公式加以改进,不需要计算顶点对之间的均值,只需将第一个顶点的坐标与水印负载一起嵌入,除第一个顶点坐标不改变外,其他顶点坐标只改变一次。在提高水印的嵌入容量和算法效率,降低数据误差,增强水印信息的安全性方面提供了解决方案。(4)提出了一种矢量地图直方图平移可逆水印算法。针对矢量地图顶点坐标之间的相关性较弱、数据存储无序的特点,设计了一种基于排序和直方图平移的矢量地图可逆水印算法,对矢量地图顶点进行排序以增强相邻顶点坐标之间的相关性,提高水印容量；利用直方图平移代替差值扩张来减小水印嵌入引入的数据误差,算法同时将原始矢量地图的HASH值随秘密信息一起嵌入,可进一步提高版权认证的能力,算法具有较高的嵌入容量和较小的数据误差,同时对坐标平移、数据排序攻击具有较强的鲁棒性。(5)提出了一种大容量矢量地图可逆水印算法。介绍了水印容量评价指标,并对差值扩张算法和直方图平移算法水印容量进行分析,差值扩张算法水印容量依赖于矢量地图顶点的可扩展差值,对于小数据量地图水印嵌入容量不高,而直方图平移算法在需要多层嵌入时产生的附加信息较多,水印潜入率不高。针对这两类算法的缺陷,设计了一种基于差值平移和奇偶检校的大容量矢量地图可逆水印算法,算法水印容量不受顶点坐标相关性的影响,有效提高了水印的嵌入容量,同时与差值扩张算法和直方图平移算法相比,算法引入的误差较小。图78幅,表11个,参考文献118篇更多还原

【Abstract】 With the constant deepening of vector map in the actual application, safety vector map copyright protection and data authentication and other issues have become increasingly prominent. As an important means of copyright protection of the vector map, digital watermarking technology has been overwhelming concerned by majority of experts and scholars, research institutions, This paper takes on the vector map lossless digital watermarking technology as the theme, studied the available vectors of vector map digital watermarking, vector map lossless watermarking framework model, in order to improve the robustness of the lossless watermarking algorithm, the watermark capacity and etc, a series of vector map lossless watermarking algorithm was studied, the main research results of this paper are summarized as follows:(1) Analysis of lossless digital watermarking techonology for vector map. Based on a thorough analysis of the vector map’s data content and data structure, a general model of lossless digital watermarking for vector map is given, Analyized the embedding position of lossless digital watermarking for vector and according to the characteristics of vector map data, studied he evaluation index of digital watermarking algorithm for vector map.(2) Study on the zero watermarking for vector map. According to the data features of the vector map, a framework model of zero watermarking for vector map is presented, analyzed the theoretical foundation of creating meaningful zero watermark. In view of the poor robustness of traditional watermarking algorithm to resist attack, according to the spatial character and the attribute character of vector map respectively, two kinds of zero watermarking scheme were designed:a zero watermarking algorithm based on feature point and block for vector map, through feature point extraction, block processing with the improved Quad tree (MQUAD), a kind of zero watermarking algorithm was designed effective against map geometric transformation (translation, rotation, scaling), data compression, vertex deletion, and map clipping; a zero watermarking scheme based on feature coding of vector map, analyzed the properties of feasibility as structural feature information using the feature code, using the stability of feature code to construct zero watermarking, which effectively improves the robustness of the algorithm against geometric transformations (translation, rotation, scaling), map generalization especially the feature editing etc.(3) The improved reversible watermarking algorithm based on difference expansion for vector map is put forward. The difference expansion algorithm is simple, has large watermarking capacity etc. Based on the traditional difference expansion algorithm, improved the calculation formula for vertex coordinates, it does not need to calculate the mean between the vertices, only the first vertex coordinates need embedding with the watermark, except the first vertex coordinates not changing, the other vertex coordinates change only once. The proposed algorithm provides a solution to improve the embedding capacity, efficiency of the algorithm, reduce the data error,and enhance the security.(4)A reversible watermarking algorithm based on histogram shifting for vector map is put forward. According to the weak correlation between the coordinates of the vertices, the chaotic characteristics of data storage, a reversible watermarking algorithm based on sorting and histogram shifting for vector maps is designed, sorting the vector map vertices in order to enhance the correlation between adjacent vertices to increase the watermark capacity; using histogram shifting instead of difference expansion to reduce the data error, at the same time, HASH value of the original vector map embedded with secret information to further improve the ability of copyright authentication, the embedding capacity algorithm has high and relatively small data error, and is robust to the coordinate translation, data reordering attack.(5) A high-capacity reversible watermarking algorithm for vector map is put forward. The evaluation index of watermark capacity was introduced, and the difference expansion algorithm and histogram shifting algorithm of watermarking capacity were analized. Difference expansion algorithm of vector map watermarking capacity depends on the vertex expandable difference, for small data map watermark capacity is not high, and the histogram shifting algorithm generated more additional information because of multi-layer embedding, which leads tolow embedding rate. Aiming at the limitation of the two kinds of algorithms, a large capacity reversible watermarking based on difference translation and parity for vector algorithm was designed, watermark capacity is not affected by the correlation of vertex coordinates, it effectively improves the watermark capacity, at the same time it bring to smaller error compared with the difference expansion algorithm and histogram shift algorithm.更多还原