Anti-Geometric Attacking Watermarking Algorithms For GF-2 Image Based On SIFT Feature Points

With the rapid development of remote sensing technology towards high spatial resolution,time resolution,spectral resolution and radiation resolution in recent years,high resolution remote sensing image has gradually become an important technical means for earth observation.GF-2 image has the characteristics of high resolution,high precision and wide application fields,and plays an important role in land and resources utilization,land planning,agriculture and marine domains.However,data leakage,piracy and illegal dissemination make it difficult to guarantee the data security of GF-2 image,and even threaten the national defense construction and national security.As a cutting-edge technology in the field of information security,digital watermarking technology has opened up a practical and feasible solution for the copyright protection of GF-2 images.It is used signal processing for digital watermarking technology to store users' information,copyright information,etc.,in the protected original data through watermark embedding,and becomes an integral part of it.Users can determine the copyright owner of the data based on the extracted watermark information,thereby providing strong technical support for the data security of GF-2 remote sensing images.Based on the data characteristics of GF-2 images and the research and discussion of common anti-geometric attack watermarking schemes,two kinds of anti-geometric attack GF-2 image digital watermarking algorithms are proposed.The main research contents and conclusions of this paper are as follows:(1)This section expounds the characteristics of GF-2 image watermarking technology and analyzes the geometric attacks.Because a scene of GF-2 image is often unable to achieve full coverage of the study area,in this case,in order to meet the needs of remote sensing mapping and theoretical research,it is usually necessary to cut and stitch the image.In addition,image processing can also lead to slight geometric distortion of image data,such as translation,rotation and deformation.Geometric attacks will lead to the destruction of the synchronization information of the watermark,and it is difficult to detect and extract the watermark.Therefore,the second chapter of this dissertation discusses the GF-2 image watermarking scheme and how to solve the problem of anti-geometric attacks.(2)In view of the fact that the traditional watermarking algorithm cannot achieve the dual functions of copyright protection and content authentication at the same time,the third chapter of this dissertation proposes a dual digital watermarking algorithm based on DWT transform and SIFT operator.The algorithm embeds two watermarks successively,which can protect the copyright of GF-2 image and realize the authentication of image content at the same time.In order to improve the imperceptibility and robustness of the watermark,firstly,the copyright information is preprocessed by Arnold as the first watermark,and the additive rule is embedded into the low frequency subband in DWT domain.Then the image feature information is extracted by SIFT operator and compressed by SVD as the second watermark.The watermark is embedded into the SVD transformed data with the first watermark,and the final image with double watermark is obtained.(3)In order to solve the problem that the existing watermarking algorithms based on SIFT feature points are not robust and can't meet the needs of copyright protection of GF-2 images due to the overlap of feature regions,the fourth chapter of this dissertation proposes a GF-2image digital watermarking algorithm based on NSCT and improved SIFT feature points.Firstly,the SIFT feature points of the GF-2 image are extracted and clustered by Mean Shift;secondly,the feature region of the image is constructed according to the key points and geometrically normalized;finally,the feature region is decomposed by NSCT,and the lowfrequency subband is selected for singular value decomposition.According to the additive rule,the singular value of the watermark information is embedded into the singular value of the lowfrequency subband,and the watermarked image is obtained by the corresponding inverse transform.