The Research On The Algorithms Of Multispectral Image And Panchromatic Image Fusion Based On Multiresolution Analysis

With the rapid development of remote sensing technology, the single band sensor has been unable to meet the needs of scientific research, so multiband sensor come into being. The multispectral remote sensing image formed by multiband sensor has high spectral resolution, but its spatial resolution is low; and corresponding panchromatic images by the formation of the single band sensor have high spatial resolution, but the spectral resolution is relatively insufficient. So how to use image fusion technology to merge these two images into a multispectral image with high spatial resolution and overcome the limitation of the single image is very important. This paper selects the fusion of multispectral and panchromatic images as the research object,according the fusion rule that it maximally improves the spatial resolution of the multispectral image and tries to reduce the spectral distortion degree of fusion image to achieve the fusion of multispectral and panchromatic image.The paper fist introduces the principle and development process of remote sensing technology, the background, application prospect, research significance of image fusion and the current situation of the fusion technology at home and abroad; then briefly describes the image fusion principle and evaluation index, summaries the current multispectral and panchromatic image fusion methods, introduces the three fusion algorithms based on the replacement of the component color space and several multiresolution analysis of the fusion algorithm in detail, evaluates these algorithm from the subjective and objective by experiments, analysis the disadvantage of these algorithms.Finally, according to the shortcomings of the existing algorithms, three kinds of fusion methods based on multiresolution are proposed. The algorithm firstly fully considering the importance of multispectral image interpolation, using interpolation algorithm based on guided filter interpose the low resolution multispectral image, getting multispectral image had the same spatial resolution with panchromatic image; then, the interposed multispectral image and panchromatic image are of the same level filtering decomposition using different filters, the multilevel filter among these filters is used to fully take into account the relationship between the fusion images in the process of filtering; and according to the different correlation coefficient, the high frequency coefficients of the replaced multispectral are judged, and the characteristic coefficients of the image imaging system are injected into the replacement process, which effectively preserves the spectral characteristics of the multispectral image; finally, it obtains fusion image by inverting the high frequency coefficients of multi spectral images obtained after the replacement and the original low frequency coefficients. Experiments show that these algorithm fully retaine the spectral information of multispectral images, and as far as possible to inject the detail information of the panchromatic image, effectively improve the effect of multispectral image fusion.