期刊名称：International Journal of Innovative Research in Science, Engineering and Technology
印刷版ISSN：2347-6710
电子版ISSN：2319-8753
出版年度：2015
期号：MULTICON
页码：627
出版社：S&S Publications
摘要：The paper presents satellite image enhancement based on SWT, NSCT with Gabor based pixel levelfusion for Pansharpening. SWT decompose the image into finer and coarser details and these details will bedecomposed into different orientation edges using NSCT. NSCT is very efficient in representing the directionalinformation and capturing intrinsic geometrical structures of the objects. It has characteristics of high resolution, shiftinvariance,and high directionality. An Image fusion involves Gabor filter to fuse finest details obtained from NSCT Itenhances the spatial information of multi spectral images, Gabor is useful to discriminate and characterize the textureof an image through frequency and orientation representation. All subbands are fused by pixel level fusion withmaximum rule to achieve better accuracy. An integration of high spatial resolution extracted from PAN images into thehigh spectral resolution of MS images generates both high spatial and spectral resolution pansharpened image. Here, agiven number of decomposition levels are used for multispectral images while a higher number of decomposition levelsare used for Pan Images relatively to the ratio of the Pan Pixel size to the MS pixel size. This preserves both spectraland spatial qualities while decreasing computation time. By applying upsampling after NSCT, structures and detailinformation of the MS images are more likely to be preserved. Hence, pansharpening is done by fusing it with detailinformation provided by the Pan image at the same fine level. The system simulated result shows that used methodprovides better resolution in these images rather than prior approaches and it also measured the performance parameterssuch as correlation, PSNR, PSD and standard deviation.
