Remote Sensing Image Fusion Based On Multiscale Transform

With the development of remote sensing technology,many kinds of sensors have been applied to data acquisition,and the available satellite images are becoming more and more abundant.In order to better meet the needs of practical application,different remote sensing image information for data analysis should effectively used and multi-source images with complementary information should be integrated.So remote sensing image fusion technology came into being.Spectral remote sensing image fusion is an important branch of remote sensing image fusion.The purpose of spectral remote sensing image fusion is to fuse multi-spectral images with high spectral resolution and low spatial resolution(MS),and to fuse panchromatic images with high spatial resolution and low spectral resolution(PAN),in order to obtain the fused images with high spatial and high spectral resolution.High-quality fused images are of great significance for subsequent land cover classification,environmental monitoring and other practical applications.In this context,this paper mainly studies remote sensing image fusion based on multi-scale transformation.Firstly,the research status of remote sensing image fusion is introduced,and the related concepts and algorithms involved in subsequent experiments provide a theoretical basis for the research work.Secondly,a multi-scale remote sensing image fusion algorithm based on IHS-NSCT transform is proposed.The intensity component of MS image is obtained by Intensity-hue-saturation(IHS)transformation of MS image;then,the intensity component of MS image and the high-low frequency component of PAN image are represented by Non-subsampled Contourlet transformation at multi-scale;after that,the low-frequency fusion rule of edge operator is proposed according to the characteristics of remote sensing image,which effectively preserves the high-low frequency component of the spectral information of the source image.Finally,the fused images are obtained by inverse transformation.Experiments show that the proposed low-frequency fusion rule effectivelyretains the spectral information of the source image,which shows that the algorithm can achieve better fusion effect than the comparison algorithm.Finally,in order to further optimize the previous algorithm,a remote sensing image fusion algorithm based on cooperative empirical wavelet transform is proposed.Empirical wavelet transform is used to construct filter banks adaptively according to the spectral characteristics of the image itself.Compared with the NSCT transform used in the previous chapter,the algorithm retains the spectral and detail information of multi-source images better,effectively improving the stability of the algorithm and achieving the optimization of the algorithm.This algorithm uses the idea of "common image" to solve the cooperative problem of to solve the cooperative problem adaptive decomposition of multi-source images.Firstly,principal component analysis is applied to multi-source image to obtain common images;secondly,IHS transform is applied to MS image and common image respectively,and empirical wavelet transform is used to obtain filter banks for intensity components of common image;thirdly,the filter banks are applied to strength components of MS image and PAN image to construct multi-scale representations respectively,and the high frequency fusion rules of local energy are used to further optimize the algorithm in Chapter 3;finally,the fused images is obtained by inverse transformation.Because of the cooperative adaptive decomposition method,the algorithm can better separate the high frequency and low frequency information of the source image,effectively improving the clarity of the remote sensing fusion image,and realizing the optimization of the algorithm.By using Quick Bird satellite data to verify the effectiveness of the algorithm,visual perception and objective evaluation criteria show that the algorithm has more advantages than other similar algorithms.This paper mainly studies remote sensing image fusion algorithm based on multi-scale transformation.According to the characteristics of remote sensing image,a remote sensing image fusion algorithm based on NSCT transform and low frequency fusion rule of edge operator is proposed.In order to further improve the fusion effect,a remote sensing image fusion algorithm based on Cooperative empirical wavelet transform is proposed,which effectively improves the quality of the fusion image.In the future,wewill further study and utilize the powerful signal processing ability of empirical wavelet transform to obtain more fusion images that meet the actual needs.