High Resolution SAR and Panchromatic Image Fusion based on Bidimensional Empirical Mode Decomposition

Bidimensional empirical mode decomposition (BEMD) method is considered advantageous for analyzing non-stationary and non-linear signals. Recently, it has been introduced in non-stationary high resolution Synthetic Aperture Radar (SAR) image processing. This letter proposes a new method for fusing high resolution Synthetic Aperture Radar (SAR) and Panchromatic images, based on BEMD method. Under this method, first, multi-resolution decomposition images are obtained, from the original images based on BEMD algorithm, the features of the original images are separated into multiple scales of spatial frequencies, called intrinsic mode functions (IMF), and then an area-based image fusion scheme is applied to fuse the images at each decomposition level. Experimental results from GaoFen-1 (GF-1) Panchromatic and COSMO-SkyMed SAR images show that the proposed method not only preserves the high spatial resolution of Panchromatic image, but also combines the target and surface information in SAR image, which are difficult to identify in the Panchromatic image. The new algorithm outperforms the wavelet transform and non-subsampled contourlet transform (NSCT), in terms of both visual effect and quantitative analysis.