Research On Second Generation Digital Watermark Algorithm Of Remote Sensing Image Based On Nonsubsampled Contourlet Transform

With the construction of smart city and the rapid development of Internet technology, the remote sensing image data, which is widely used in territorial planning,mapping,navigation,military and other fields,shows its broad application prospects.At present, the rapid development of network technology makes the remote sensing image data transmission and download become more convenient, but the follow-on issues of data security become further strengthened.Digital watermarking technology, as an important technology in data security, is an effective way to solve the problem of remote sensing image data security. Digital watermarking technology can provide effective method for copyright protection, user tracking and so on, by embedding watermark information into the remote sensing images, such as copyright and userinfo. Thus it can provide the powerful technical support for the safety use of remote sensing image data.With the constant-depth study by the domestic and oversea researchers, digital watermarking technology has been developed from the first generation watermarking technology to the second generation watermarking technology. In the second generation digital watermarking algorithm of remote sensing image, the watermark information is embedded not only by changing the global pixel value of the image or a certain transform coefficient,but also using some important features of remote sensing image,which makes information and remote sensing image data more closely linked. Thus,this method greatly improves the robustness of the watermarking algorithm.In this paper, the digital watermarking algorithm for remote sensing image has been studied, based on the second generation digital watermarking technology. The main research work is as follows:(1) Based on the characteristics of remote sensing image,the characteristics and requirements of digital watermarking technology for remote sensing image are expounded. Through analyzing the main procedure of remote sensing image data processing,the main methods of watermark attack are concluded,which remote sensing image will face in the productive application. Thus,the robustness of digital watermarking algorithm for remote sensing image is improved specifically.(2) The principle and characteristics of nonsubsampled contourlet transform(NSCT) are analyzed. By means of the theoretical analysis and simulation experiments,the method to choose embedding position of the watermark in the NSCT domain is studied. The foundation for the research on digital watermarking algorithm for remote sensing image based on NSCT is laid.(3) By the analysis of the characteristics of grayscale adjustment and the principle of Mean Shift algorithm, a second generation watermarking algorithm resisting to grayscale adjustment of remote sensing image based on NSCT is proposed. The algorithm extracts the strong texture region by Mean Shift clustering with the texture features extracted from gray-level co-occurrence matrix, and then adaptively embeds watermark into the middle frequency subband image in NSCT domain. It solves the problem of being unable to resist grayscale adjustment attack, and also has good robustness to some conventional image processing and cropping attacks.(4) A robust watermarking algorithm for remote sensing image based on feature points is proposed by combining SIFT feature points and image normalization technique. It extracts feature regions from the original remote sensing image based on the SIFT feature points, and then make the feature regions normalized. The watermark information is embedded into the coefficients of NSCT by odd-even quantization method. The experimental results show that the proposed algorithm is robust to some conventional image processing and geometric attacks, such as translation, scaling,cropping and so on.