MS-Pansharpening Algorithm Based on Dual Constraint Guided Filtering

The difference and complementarity of spatial spectral information between multispectral (MS) image panchromatic (PAN) have laid the foundation for fusion two types images. In recent years, MS PAN (also known as MS-Pansharpening) has gained attention an important research area in remote sensing (RS) processing. This paper proposes MS-Pansharpening algorithm based on dual constraint Guided Filtering nonsubsampled shearlet transform (NSST) domain. innovation is threefold. First, guided filtering (DCGIF) model, region average gradient correlation vector formed by neighborhood elements proposed. Further, detail extraction scheme, provided, which extracts more complete accurate information, thus avoiding, to some extent, distortion caused injection non-adaptive information. Second, weighted preservation band spectra, model determines weight each pixel proportion bands original image, ensures fused image. Finally, a new NSST domain high frequency sub-bands are used extract effective details. Then proposed DCGIF through joint method regional energy matrix. inject it into fusion. Experimental results show that approach better effect than conventional algorithms, can effectively improve resolution maintain characteristics MS.