Research On Super-resolution Image Fusion Algorithm Based On Multi-spectral

Remote sensing technology and multi-sensor technology develop continuously, making the ability of acquisition and processing remote sensing data enhanced constantly.By means of multi-sensor technology, people can get the image information of different spectra, different scales, different time, different phase from the same area or the same target.Through post-processing of image data, image fusion compensate for the shortage of image information obtained by a single sensor,fusing the same scene of multiple images to get more abundant,accurate,reliable information.Compared to a single source image, the image fusion improve the image definition, reduce the ambiguity,incompleteness,uncertainty and error.Multi-spectral image in remote sensing images has abundant spectrum information but low spatial resolution, full-color image has high spatial resolution, but does not reflect the spectrum information. In practical applications, to obtain both high spatial and spectral resolution image costs too much,so it is often through the subsequent image processing techniques to fuse multi-spectral image and full-color image,making sure that the image fusion own both high spatial resolution which full-color image has,and also the spectral information from the fine description of target features,which is characteristic of multi-spectral images.According to the characteristics of multi-spectral image and full-color image, this paper analyze several classical fusion algorithm of the pixel-level image,which including IHS,PCA,Brovery,and Multi-resolution fusion algorithms and so on.The traditional algorithm such as HIS transform, PCA transform is simple, but there is a certain spectral distortion in the image fusion;and multi-resolution image fusion algorithm,often increase the fusion effects at the expense of the algorithm complexity.This paper propose a fusion algorithm based on compressed sensing, revise the fusion results of traditional PCA changes and reduce the spectral distortion, while retaining its rich details.This paper selects several classical fusion algorithms for the simulation experiments, and get the fusion results. Accroding to subjective and objective evaluation criteria,we compare the fusion results,to demonstrate the algorithm is feasible Experiments show that the proposed algorithm can achieve better fusion effect.