支持版权保护和数据验证的地理数据库水印方法研究

地理数据库是一种重要的数据资源。地理数据库所有者和使用者作为数据资源的供给方和需求方，他们关注的数据利益有所不同，前者关注数字版权保护，而后者更关注数据完整性和正确来源，本项目拟研究支持版权保护和数据验证的地理数据库水印方法，同时保护两者的权益。通过分析水印攻击机理，提出"水印同步、局部嵌入、重复嵌入"三个地理数据水印算法设计原则，研究同时抵抗常见的三种攻击的私有地理数据库水印方法用于保护地理数据库版权；提出基于"两次水印生成两类公钥"的公有地理数据库水印方法，不但可实现公开、重复验证地理数据库的完整性和数据来源，而且对于被篡改的数据可实现篡改定位以及篡改类型识别；并在此基础上提出两阶段水印集成方法，将私有水印和公有水印同时隐藏在地理数据库产品中，这样在水印化地理数据库被发布之后，数据库使用者能够公开、重复验证该数据库的完整性，数据库所有者也可通过私有水印检测算法声明所有权归属。

Geographical database is one kind of the important data resources. As data resource provider and requester, the data owners and end-users of geographical database have different profits relevant to data resource. The owners of geographical database focus on digital copyright protection, whereas the end-users focus on data integrity and original source. This project studies the geographical database watermarking method which provides both copyright protection and data authentication to safeguard the interests of data owners and end-users at the same time. By analyzing the mechanisms of typical watermark attacks of geographical database watermarking schemes, we sum up three principles for designing robust geographical database watermarking schemes, which are 'watermark synchronization', 'locally embed' and 'repeatedly embed', and design robust private geographical database watermarking scheme which is resilient to three kinds of typical watermark attack ways following these three principles to protect the copyright of geographical database. Based on the idea of "twice watermarking generate two public keys", we propose a public geographical database watermarking scheme, which can be used not only to verify the integrity and original source of geographical database openly and repeatedly, but also to locate the tamper of the data and identify the type of tamper. Based on private and public watermarking scheme, we study the method of watermarking integration, which hides private watermark and public watermark into geographical database. After the watermarked data is released, end-users can verify data integrity openly and repeatedly, and data owners can declaim ownership by detecting private watermark.