Research On Fusion Technology Of Sar Optical Remote Sensing Image

Remote sensing satellites provide people with extremely rich remote sensing image data of different spatial resolutions and different spectra.Massive remote sensing data have made great contributions to the fields of military,natural disasters,surveying and mapping,aerospace and other fields.Due to the different imaging principles of different sensor types and other factors,the imaging of a single sensor cannot present all the information in the scene,resulting in the imaging results of multi-source sensors for the same area often in geometric characteristics,spectral information,spatial resolution.There are huge differences and limitations in terms of rate,gray distribution,etc.Multisource remote sensing image fusion combines the spatial consistency and information complementarity of different sensor images of the same scene,eliminates redundant parts in the image,and fuses the advantageous information in the two images to obtain a high spatial resolution,Fusion images with hyperspectral information,clearer geometric features and more comprehensive details.Nowadays,how to efficiently combine the high spatial resolution characteristics of Synthetic Aperture Rader（SAR）and the hyperspectral information in light images to obtain a more comprehensive fusion image is a multisource remote sensing image fusion technology.research focus in the field.Therefore,my research result of the thesis is as follows:（1）Focusing on image fusion,discuss its research background and significance,and introduce its status in this world;from the imaging mechanism,summarize the characteristics of visible and SAR pictures,and then summarize the picture process before fusion;The shortcomings of traditional algorithms and the direction of improvement are introduced from three aspects: level and decision level;Finally,I introduce the evaluation of fusion image results in detail.（2）A fusion algorithm based on fast wavelet transform is proposed for images of areas where the SAR echo is similar to specular reflection,such as oceans and airport runways.In view of the large dark area in the background and the fact that the target area we pay attention to often has the characteristics of corners,edges,etc.and the gray level of the picture is very unusual,I design two rules for high and low part of picture.Comparison by experiment,the algorithm retains more spectral information,and further make the image details and edges more delicate and geometric characteristics more prominent.（3）For the images of cities,counties and towns with a large number of strong reflectors,we propose an NSCT fusion algorithm.in view of the characteristics of large energy and frequent grayscale transformation in this area,I design different frequency fusion rules in high and low part of picture.Comparison by experiment,the fusion image of this algorithm has higher contrast and clearer image,and at the same time highlights the salient features of the source image.