Research And Development On Spatial Data Watermarking Technology

The rapid development of computer technology and Internet techniques makes the distribution of digital contents easier than ever. Correspondingly, it also becomes harder to protect the digital copyrights of various digital medias. As a new technology developed in recent ten years, digtial watermarking which adds structure called watermark to the digital data perceptibly or imperceptibly to assert the ownership or to verify the data integrity provides potential solutions to the issue. the watermarking for some general multimedia data, such as digital images, audio, videos and 3D models, has been given much more attention than other data types wheres the spatial data watermarking research has been just started and the related works are relatively much fewer.After making a in-depth and effective exploration of the GIS spatial data watermarking, this paper analyzes the characteristics of spatial vector data watermarking; proposes different algorithms for Linear features, different types of area features and DEM data respectively; analyzes and tests the robustness and security of the algorithms; in the end, proposes a spatial data watermarking protocol based on an authoritative and trusted third party, making an analysis and simulation of it. Specifically, the content of this research and innovation mainly include the following:1. The evaluation of spatial vector data watermarkingBy making a study on the simple object model, spatial vector graphic attributes and features, and the spatial relations, this paper analyzes the characteristics of spatial vector data which have much association with watermarking and summarizes all the possible feature set which can be used as watermarking cover data.We point out that the uncertainty of spatial vector data has provided hiding places for watermark bits. The basic principles that the watermarking must follow on is that the distortions induced by watermarking must lie within the range of data uncertainty. A data usability model and fidelity model are established respectively, allowing the quantitative evaluations of watermarking performance.All the possible threatening attacks for the spatial vector data watermarking are proposed and classified.Additionally, an evaluation model for spatial vector data watermarking algorithms is built based on the data usability model and data fidelity model.2. Robust linear feature watermarking technologyDifferent watermarking mechanisms based on the graphic features the data shows and the geo-spatial meanings the data represents, have been proposed for watermarking linear features and area features respectively. A new watermarking algorithm(denoted as LFWA) for linear features has been constructed. In the algorithm, a robust ID has been defined to label object and a shape describing parameter has also been defined which is used as watermark cover data. Watermark bit is embedded by adjusting the distribution pattern of the cover data.3. Robust area feature watermarking mechanismsA metric measure for disjoint relations is introduced. Based on the disjoint relations, a new algorithm for watermarking polygons (denoted as PWDR)have been designed. Firsty, the statistics of disjoint relation metric measures between the objects within the same data patch are extracted and used as watermark cover data. Then the watermark cover data is divided evenly into several subsets according to a keyed HASH function. A watermark bit is inserted into a subset by adjusting the distribution pattern of the cover data within the subset. Testing results show that the PWDR algorithm is very robust. However, the time cost of the program is much high and the watermark capacity is relatively small. To overcome the shortcomings, some simplifications and improvements have been made to the PWDR algorithm(the modified PWDR algorithm denoted as S_PWDR).In the S_PWDR algorithm, the polygons are sorted by a cryptographic keyed hash function. Then the sorted polygons are divided evenly into several subsets by assigning a set of adjacent items into one subset. Pairs are made using every two adjacent polygons. For each pair, disjoint relation metric measures are computed by reference each other within the pair. A watermark bit is embedded into subset by adjusting the distribution pattern of the metric measures of the subset.For the types of vector data such as administrative map, a new algorithm based on the feature points of polygon boundaries is proposed. In this algorithm, polygons are treated as multi-line strings and then watermarked using the algortithm for linear features.4. A watermarking method for DEM dataFrom the viewpoint of the characteristics of DEM data and the geographic features of a surface terrain conveyed by the DEM data, a watermarking method(denoted as DEMWCC) based on the critical contours is proposed for watermarking DEM data. In the algorithm, critical contours which are essential for conveying the topographic features of a terrain or for some specific applications are firstly extracted from the DEM grid data. The LFWA algorithm is extended to watermark the critical contours. Then a DEM grid data is re-constructed to incorporate the watermarked contours. When detecting the watermark, the watermarked critical contours are recovered first, and then the detection algorithm of LFWA is employed to extract the watermark messages in the contours.5. A watermark security protocol for spatial dataBased on the above watermarking algorithms for spatial data, a watermark security protocol have been built, which is based on a Authoritative Trusted Third Party (ATTP). The protocol can protect the interests of both the copyright owners and the consumers. The security of the protocol is analyzed and a prototype system implementing the protocol has also been built in the end.