Abstract: Non-Subsampled Contourlet Transform (NSCT) is a improvement of Contourlet Transform (CT). The important properties of the Non-Subsampled Contourlet Transforms are shift-invariance, high resolution, and high directionality. Non-Subsampled Transform represents high directional information and captures geometrical structures of the objects. In the proposed method the given numbers of decomposition levels are applied to multispectral Image and the Panchromatic Image in the relative ratio. Fine levels and coarse levels are obtained. During the Simulation process Bilinear Interpolation method is applied for fine levels of the Panchromatic Image and Multispectral Image. The coarse level of multispectral Image given to a local edge preserving filter, and later inverse NSCT is applied to get the Pan-Sharpened Image. Non-Subsampled Contourlet Transform preserves spectral and spatial information. The computation time is decreased. Considerably the detailed information can be preserved by applying Up sampling after NSCT. By this Upsampling after NSCT, the coefficients are fused by the application of Fusion Rule. After Inverse NSCT, the obtained Image is a Pan Sharpened Image. An additional Local Edge Preserving Filter (LEP) is used before NSCT to get Pan Sharpened Image with high Spatial and Spectral qualities. By Images and quality indices values, we can observe the improved Spatial and Spectral qualities.
Keywords: NSCT, Bilinear interpolation, Image fusion, local edge preserving filter (LEP), Pan-sharpened Image or fused image.