Research On Authentication Methods For Vector Geographic Data Based On Fragile Watermarking

Vector geographic data, as an important part of fundamental geospatial data, is strongly related to national security. It also plays a significant role in the development of national economy and society. Come as the age of the Internet, the processes of distributing, transmitting, storing and applying data rise along the problem of authenticity and integrity which has an influence on security, application and sharing of vector geographic data. How to ensure the authenticity, integrity and reliability of vector geographic data with reliable technologies becomes the focus point of studies in field of geographic information security.Fragile digital watermarking is a cutting-edge technique in the field of information integrity authentication. The integrity of transmitted information can be verified effectively and the change of data can be located by using this technique. Fragile watermarking is an effective remedy for detects of digital signature and provides effective solutions for integrity authentication for vector geographic data. The research on integrity authentication for vector geographic based on fragile watermarking is vital.Based on essential features of vector geographic data, the dissertation conducts systematic research on integrity authentication for vector geographic data. Fragile watermarking and semi-fragile watermarking are applied on the process of integrity authentication in this dissertation. The main research results are as follows:(1) Distinctions between vector geographic data and other types of data are described from various aspects of data collection, data storage and data applications. Technical characteristics of fragile watermarking for vector geographic data are summarized for the purpose of laying the foundation of constructing the model of integrity authentication for vector geographic data and designing the algorithms.(2) On the basis of analyzing existing extract authentication algorithms for vector geographic data, an extract authentication algorithm for vector geographic data based on a neo fragile watermarking algorithm is proposed in the dissertation. By sorting points and embedding watermark information produced by the last point into the next point, relations between points are built. Deleting a point will destroy relations between points so that the deleted point can be detected when conducting extract authentication. To increase the positioning accuracy and precision of this algorithm, the points are blocked into several parts with the same area before sorting and embedding. Considering that vector geographic data are unevenly distributed sometimes, the furthermore improvement of the algorithm is to determine the area of blocks by limiting the number of points in one block. The improved algorithm is more flexible and fast.(3) According to the features of the light field method for data compression and the characteristic of semi-fragile watermarking, a selective authentication algorithm for vector geographic data is proposed. A new anti-compression semi-fragile watermarking is used in the algorithm. Before embedding, points are divided into feature points and non-feature points by the light field method for data compression. Then the algorithm uses different watermarking methods on feature points and non-feature points so that the authentication for vector geographic data can be selective. The experiment shows that the proposed algorithm is robust to resist the attack of the light field method for data compression and other types of modification can be detected by the algorithm. Therefore the proposed algorithm achieves good effect on selective authentication.