Many proposed image watermarking techniques are sensitive to affine transforms, such as rotation, scaling and translation. In this paper, a localized affine transform resistant watermarking is designed utilizing Krawtchouk transform and dual channel detection. Watermark is inserted into the significant Krawtchouk invariant moment. Watermarking based on Krawtchouk moments is local, which permits to the watermark to be embedded at the most significant information-wise portion. Watermark embedding intensity is modified according to the results of performance analysis. The convergence of closed loop embedding system is proved. An optimum watermark detector is designed with the introduction of dual channel detection utilizing high order spectra detection and likelihood detection. The detector extracts watermark blindly utilizing Independent Component Analysis. The computational aspects of the proposed watermarking are discussed. Experimental results demonstrate that this watermarking is robust with respect to attacks produced by watermark benchmark—Stirmark.