Research On Technique For Automatic Watermark Detection Of Vector Geographic Data

Digital watermarking is an effective solution of copyright protection of Vector geographic data. However, the research on digital watermark detection of Vector geographic data is still weak, and the system theory and effective detection method has not yet been established. Furthermore, the watermark was detected or derived to make certain the ownership of the suspected pirate data only after the data was found, and this affected the application of digital watermarking technology seriously. In view of the current status and unique needs of Vector geographic data watermark detection , this dissertation proposes an independent watermark detecting method based on mobile agent to find the pirate data out of the huge amount of vector data distributed over the internet automatically. This study is supported by The HI-TECH Research and Development Program of China (863 Program) ~ Research on Key technique for Copyright Protection of GIS Vector Data (NO.2006AA12Z222).This dissertation includes three main sections to discuss the key problem: blind watermark detection of Vector geographic data, independent watermark detecting method and the security of watermark agent. The major contents of studying of this dissertation are as follows:(1) By analyzing the technique of watermark detection, this dissertation presents 4 watermark detection algorithms of Vector geographic data including spatial algorithm based on statistic variance, algorithm in DWT transform domain based on quantification and algorithm in DFT transform domain based on multiplication in count of the character of the Vector geographic data signal.(2) This dissertation presents an independent watermark detecting method based on Mobile Agent and the framework for watermark detection prototype system has been developed and realized.(3) In order to satisfy the security and privacy requirement of host and watermark agent, this dissertation describes how to apply encryption and digital signature to watermark agent and host.Considering multiple watermark involved uniform threshold and statistical distribution of Vector geographic data, normalization of Correlation-Detection is not suitable for DFT blind watermark detection. Experiments show that the spatial algorithm based on statistic variance is more suitable for application because of robustness and low distoration. Furthermore, the watermark detecting method based on mobile agent this dissertation presented is a viable solution for independent watermark detecting, and encryption and digital signature can improve the security of watermark agent.The innovations of this dissertation mainly include the method of watermark detection of Vector geographic data and the method of initiative watermark detection.