Research On Remote Sensing Image Fusion Based On Variational Methods

In recent years,remote sensing images have been widely used in various fields of domestic economy.However,due to the limitation of hardware,satellites sensors usually catch one-band panchromatic images with high resolution and multispectral images with low resolution.Hence,image fusion has been introduced to the field of remote sensing,fusing low-resolution multispectral images and panchromatic images for the same region into one high-resolution target image,which is utilized in further analysis and process.This process is called pan-sharpening.The key point of variational pan-sharpening methods is creating an energy functional based on assumptions regarding the relationship between the input images and the target images,making the pan-sharpening problem an optimization problem.At present,compared with conventional methods,variational pan-sharpening methods have positive results in preserving spatial and spectral information but still have some defects,such as the lack of spatial information and high time complexity.Aiming at improve the performance of variational pan-sharpening methods,the research content of this thesis are as follows:Existing variational pan-sharpening methods typically consider gradient as the descriptor of the spatial information and enforce it to be consistent between the panchromatic image and fused image because of their shared scenes.However,the gradient feature is not sufficient to exploit the abundant spatial details.Second-order derivative,which is balanced with the first-order derivative,is introduced to build a variational model such that additional geographic information is extracted from the PAN image.Then corresponding numerical schemes to determine the minimal solution of this model are provided.Through experiments on data from four different satellites,the proposed method is proven to enhance the spatial quality of the fused result and demonstrate stable performance in preserving the spectral information.Because current variational pan-sharpening methods usually have high time complexity,a method combined with principal component analysis is proposed in this thesis.Existing variational pan-sharpening methods typically fuse panchromatic image and a band of multispectral image individually,while proposed method acquiring the first principal component by PCA decomposition and this component will be fused with panchromatic image.The first principal component includes major information of multispectral image,but has a size of one-band image.Hence,the proposed method could reduce the time consumption in fusion process while keeping producing result images with high quality.