关键词: 遥感图像, 局部化水印, Harris-Laplace, 图像归一化, Contourlet

Abstract: Against the problem of existing watermarking algorithms for remote sensing image’s poor resistance to shear, rotate, horizontal mirror, and combined geometric attacks, it proposes a new localized watermarking algorithm for remote sensing image. It scrambles the watermark image by Arnold, then scans the scrambled image into one-dimensional vector as watermark information; it extracts Harris-Laplace feature points in the host remote sensing image, and determines feature regions for watermark embedding according to the points; Finally, the selected feature region is specified and normalized, and it is applied Contourlet transformation, watermark is embeded into the selected transform coefficients. Experiments have shown that the algorithm has good robustness against conventional signal processing such as filtering, noise, JPEG compression etc., and it can better resist to geometric attacks such as rotation, scaling, moving and their combination.