矢量居民地群零水印算法研究

发布时间：2018-10-30 15:07

【摘要】：GIS矢量地图数据是地理空间数据的重要组成部分,是一种重要的国家基础地理信息产品,被广泛应用到了社会各个行业。然而矢量地图数据的数字化存储方式在为传播、复制提供便利的同时,也给矢量地图数据版权安全问题带来了隐患。数字水印技术是一种版权判定与盗版追踪的新型技术,经过大量研究和实验,被成功应用在空间数据版权保护问题上。居民地是地图表达中重要的、不可或缺的地图要素,其在城市规划、政府决策分析、学术研究等领域具有重要应用价值。而矢量居民地数据在传播、复制过程中难免会被窃取、盗版。因此,迫切需要研究一种保护居民地数据版权的方法。目前,有关矢量地图数据的数字水印算法研究成果丰硕,而将现有这些水印算法移植到居民地数据上可能会损坏居民地数据精度或导致要素间拓扑关系的变化。鉴于此,本论文利用矢量零水印技术,结合矢量居民地数据特征及现有水印算法缺陷,对矢量居民地数据零水印算法进行了系统研究。本论文取得以下一些研究成果及创新点:(1)考虑到传统嵌入水印算法存在水印嵌入导致原始数据精度损失或拓扑关系的变化等不足,提出用零水印思想设计针对居民地数据的水印构造算法。(2)在分析矢量居民地图形特征的基础上,提出一种基于顶点矢量的居民地零水印算法。其设计思路是:首先,利用Logistic混沌置乱方法对原始水印图像做加密处理;然后,计算居民地多边形的顶点矢量,并将所有居民地的顶点矢量映射成一个特征矩阵;最后,将该特征矩阵与置乱后的水印图像作异或运算实现零水印构造。(3)针对(2)中提出的方法存在的不足,提出一种基于不变距和Hilbert码的矢量居民地零水印算法。其设计思路是:为增强零水印的抗裁剪能力,算法利用Hilbert编码对数据分块以便构造多个零水印。每块零水印的构造方法为:首先,利用Arnold技术置乱原始水印图像;然后,计算居民地图形的矢量不变矩集并将其映射成一个二值矩阵;最后,将置乱后的水印图像与不变矩集所映射的二值矩阵进行异或运算,完成零水印构造。本论文分别对提出的上述两种算法进行了鲁棒性检验实验,实验结果表明,基于顶点矢量的居民地零水印算法能够有效抵抗几何攻击以及一定程度的裁剪攻击;基于不变距和Hilbert码的矢量居民地零水印算法对几何攻击、一定程度裁剪攻击以及10%以内的增、删点攻击等均具有良好的抵抗能力。
[Abstract]:GIS vector map data is an important part of geospatial data, is an important national basic geographic information products, has been widely used in various sectors of society. However, the digital storage of vector map data not only facilitates the transmission and reproduction of vector map data, but also brings hidden trouble to the copyright security of vector map data. Digital watermarking is a new technology of copyright determination and piracy tracing. After a lot of research and experiments, it has been successfully applied to the protection of spatial data copyright. Residential land is an important and indispensable map element in map expression. It has important application value in urban planning, government decision analysis, academic research and so on. And vector residents in the dissemination of data, copying process will inevitably be stolen, piracy. Therefore, there is an urgent need to study a method to protect the copyright of residential data. At present, great achievements have been made in digital watermarking algorithms for vector map data. However, transplanting these watermarking algorithms to residential data may damage the precision of the data or cause the change of topological relationship between the elements. In view of this, this paper makes use of the vector zero watermark technology, combines the vector resident data characteristic and the existing watermark algorithm flaw, carries on the systematic research to the vector resident land data zero watermark algorithm. In this paper, the following research achievements and innovations have been achieved: (1) considering the shortcomings of the traditional watermark embedding algorithm, such as the loss of the original data precision or the change of the topological relationship due to the watermark embedding, In this paper, we propose a watermarking algorithm based on the idea of zero watermark. (2) based on the analysis of the shape characteristics of vector resident map, a new algorithm based on vertex vector is proposed. The design ideas are as follows: firstly, the original watermark image is encrypted by Logistic chaos scrambling method, then the vertex vector of resident polygon is calculated, and the vertex vector of all resident land is mapped into a characteristic matrix. Finally, the feature matrix is constructed by using XOR operation with the scrambled watermark image. (3) aiming at the shortcomings of the proposed method in (2), a vector resident zero watermarking algorithm based on invariant distance and Hilbert code is proposed. The design idea is: in order to enhance the anti-clipping ability of zero-watermark, the algorithm uses Hilbert coding to divide the data into blocks in order to construct multiple zero-watermark. The method of constructing every zero watermark is as follows: firstly, the original watermark image is scrambled by Arnold technology, then the vector invariant moment set of the resident map is calculated and mapped into a binary matrix. Finally, the scrambled watermark image and the binary matrix mapped by the invariant moment set are computed by XOR, and the zero watermark is constructed. In this paper, the robustness tests of the above two algorithms are carried out, and the experimental results show that the vertex vector based resident ground zero watermarking algorithm can effectively resist geometric attacks and some degree of clipping attacks. The vector resident zero watermarking algorithm based on invariant distance and Hilbert code can resist geometric attacks, clipping attacks to a certain extent and increasing and deleting attacks by less than 10%.
【学位授予单位】：兰州交通大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：P208

【参考文献】