Authentication Scheme Based on Principal Component Analysis for Satellite Images

This paper presents a multi-band wavelet image content authentication scheme for satellite images by incorporating the principal component analysis (PCA). The proposed schemeachieves higher perceptual transparency and stronger robustness. Specifically, the developed watermarking scheme can successfully resist common signal processing such as JPEG compression and geometric distortions such as cropping. In addition, the proposed scheme can be parameterized, thus resulting in more security. That is, an attacker may not be able to extract the embedded watermark if the attacker does not know the parameter.In an order to meet these requirements, the host image is transformed to YIQ to decrease the correlation between different bands, Then Multi-band Wavelet transform (M-WT) is applied to each channel separately obtaining one approximate sub band and fifteen detail sub bands. PCA is then applied to the coefficients corresponding to the same spatial location in all detail sub bands. The last principle component band represents an excellent domain forinserting the water mark since it represents lowest correlated features in high frequency area of host image.One of the most important aspects of satellite images is spectral signature, the behavior of different features in different spectral bands, the results of proposed algorithm shows that the spectral stamp for different features doesn't tainted after inserting the watermark.