A geometrically robust image watermarking algorithm(GWPZM)is proposed by using the rotation invariant property of pseudo-Zernike moments.The pseudo-Zernike moments of the cover image are first computed.Those suitable for watermark embedding are selected and tested according to their magnitude changes after being modified.Optimal moment set is determined and the moments within it are modified according to the random watermark bits.The difference vector is reconstructed and added to the cover image in spatial domain to obtain the watermarked image.In detection,the pseudo-Zernike moments of the possibly attacked images are computed and the watermark is extracted by comparing the magnitudes of selected moments with the optimal moment set as the key.Experimental results show its robustness to rotation,scaling,combined attacks,as well as common attackes such as JPEG compression and added noise.