【作者】 任娜；

【导师】 刘学军； 朱长青；

【作者基本信息】 南京师范大学 ， 地图学与地理信息系统， 2011， 博士

【摘要】 遥感影像数据是国家基础设施建设和地球科学研究的支撑性成果之一,在国家经济、国防建设中占有十分重要的地位。随着遥感影像数据生产和应用进程的不断加快,遥感影像数据的安全和版权问题也日益凸显,侵权、泄密、非法传播与使用等行为屡禁不止,仅依靠现行的安全保密措施,根本无法有效阻止与打击各种非法行为。数字水印技术作为一种崭新的信息安全技术,为遥感影像的安全保护提供了一种切实可行的解决途径。数字水印技术在数据中嵌入秘密水印信息,成为源数据不可分离的一部分,由此来确定版权拥有者、跟踪侵权行为、认证数字内容等,从而能够为遥感影像数据的安全使用提供强有力的技术支撑。本文对遥感影像水印技术的模型、算法进行了研究,主要研究工作和成果如下：(1)基于遥感影像自身的特征,阐述了遥感影像数字水印技术的特点和要求。分析了遥感影像几何攻击类型,探讨了不同几何攻击形式对遥感影像数字水印算法的影响。(2)基于对遥感影像特点的分析,提出了一种基于密钥矩阵和模板匹配的抗几何攻击的遥感影像水印算法。在不改变影像最高有效位的前提下,采用密钥矩阵嵌入水印信息,满足了遥感影像嵌入水印后数据的可用性。在水印提取过程中,采取模板匹配方法,提高了水印提取的正确性。(3)基于对遥感影像几何攻击的分析,提出了一种基于映射机制的遥感影像水印算法。通过影像数据与水印信息建立的多对一函数关系,很好地确立了水印的嵌入位置,并将一位水印信息同时嵌入到多个影像数据中。针对不同波段的遥感影像,选择了不同的映射参数。采用映射机制确定水印嵌入位置的方法,有效地解决了水印信息不同步的问题,提高了算法的鲁棒性。(4)分析了DCT变换应用于遥感影像水印中的优势,提出了一种基于伪随机序列和分块DCT的遥感影像水印算法。通过伪随机序列选取嵌入块位置,增加了水印算法的安全性,并提高了算法的效率。基于已构建的水印算法,提出了两种改进算法,能够有效抵抗遥感影像的复合攻击。(5)针对遥感影像数字镶嵌等处理涉及到的拼接和色调调整等,对抗拼接和色调调整的遥感影像数字水印进行了研究,提出了基于黄金分割法搜索的水印检测算法,提出了基于线性特性的水印检测算法,提出了基于动态量化的水印嵌入和检测算法。算法能有效抵抗拼接或色调调整攻击。更多还原

【Abstract】 Remote sensing data is one of the supportive achievements in the field of national infrastructure and earth science research, and plays a significant role in national economy and defenses. With the accelerating process of the production and application, an important issue that arises in the security and copyright issues of remote sensing data is becoming increasingly salient. It has been recognized for quite some time that current security laws and technologies are inadequate for dealing with the behavior of infringement, disclosure, illegal dissemination and usage.Digital watermarking technology is a new information security technology, and can provide a feasible and practical solution for the copyright protection for the remote sensing image. Digital watermarking is the process of embedding information into digital data such that the information can later be extracted or detected for a variety of purposes including copyright prevention, ownership authentication, infringement tracking and tamper detection. Digital watermarking is being deemed essential to provide a strong technical support to legitimate use of remote sensing data.The digital watermarking models and algorithms for remote sensing image are studied. The main results are as follows:(1) Based on the own characteristics of the remote sensing image, the characteristics and requirements of the digital watermarking technology for remote sensing image are outlined, the types of geometric attacking of remote sensing image are analyzed, and the effects of different geometric attacking on the impact of digital watermarking for remote sensing image are explored.(2) Based on the analysis of the characteristics of remote sensing image, an algorithm based on key matrix and template matching to resisting geometrical distortions is proposed. In the premise of not changing the image in the most significant bit, key matrix is used to embed the watermark, so that to meet the availability of watermarked remote sensing image. In the watermark extraction process, the template matching method is used to improve the watermark extraction correctness.(3) Based on the analysis of geometrical distortion of remote sensing image, an algorithm based on mapping mechanism is presented. To locate the embedded mark, the one watermark information is embedded in more than one pixel value using the function established by image data and watermark. For the different bands of the remote sensing images, different mapping parameters are chosen. The problem of synchronization can be resolved by using the mapping mechanism to determine the location of embedded watermark. The robustness of the presented algorithm is also improved.(4) The advantages of the digital watermark based on DCT for remote sensing image are analyzed. An algorithm based on pseudo-random sequence and block DCT is proposed. The embedded block location is selected by the mechanism of pseudo-random sequence, so as to increasing the security, and improving the efficiency of the proposed algorithm. Based on proposed algorithms, two improved algorithms are presented. These algorithms effectively resist the combined attack of remote sensing images.(5) Taking into account the splice and color adjustment involving the digital mosaic process, the algorithms to resist the splice and color adjustment are studied. There are three algorithms, including the detecting algorithm based on search mechanism of golden section method, the detecting algorithm based on the feature of linearity, the detecting algorithm based on the mechanism of dynamic quantitative. These proposed algorithms can effectively resist the splice or color adjustment.更多还原