Research On Methods Of SAR And Optical Remote Sensing Image Fusion

With the progress of the times,scientific research in the field of remote sensing has gradually matured,and humans can easily obtain a large amount of image data through various remote sensing satellites.SAR and visible light imaging have a very wide range of applications.Although the gray values ??of the two differ greatly,they have their own characteristics.The imaging mechanism of visible light image is light reflection imaging,which depends on the light source,so it has a large amount of spectral information and detailed information,and the intuitive effect is very good.The imaging mechanism of SAR images is active microwave imaging,which is more sensitive to buildings such as bridges and houses.It has rich edge contour and structural information.Due to the limitations of hardware sensors,remote sensing images cannot have both high spectral and high spatial resolution characteristics.With the increasing application fields and requirements of remote sensing images,obtaining high-quality fusion images with the advantages of both SAR and visible light images has always been the focus of the current fusion technology field.Based on the analysis of the imaging mechanism,development trajectory and current research status of SAR and visible light images,this thesis fully summarizes the current status of image fusion research at home and abroad.On this basis,an improved fusion algorithm for SAR and visible light images is proposed.Main content as follows:1.Propose an improved fusion algorithm combining PCNN and NSST-HIS.This method uses the computational efficiency and translation invariance of NSST and HIS to decompose SAR and visible light images,and uses PCNN model adaptive fusion rules for low-frequency component fusion,which can retain more spectral information and effectively reduce spectral distortion.The target features reflect the details of the image to a greater extent;the New Sum of Modified Laplacian fusion rule is used on the high-frequency subband components.The quality of the fused image obtained by this method has been significantly improved.2.Propose a model fusion algorithm based on dual-channel PCNN.This method takes advantage of the dual-channel PCNN model to maintain the integrity of image edge feature information,formulates corresponding low-frequency and high-frequency fusion rules,and obtains the fusion coefficient excitation model,so as to obtain a fused image with clearer edge feature information and contours,and suppress the emergence of the pseudo-Gibbs phenomenon.The phenomenon of unclear edge contour features of the fused image has been significantly improved.The paper uses the visible light image captured by Landsat-8 satellite and the SAR image captured by the Sentinel 1 radar satellite as the basic data,and compares the fusion experiment to illustrate.The experiment proves that the method proposed in this paper compares the traditional fusion method from visual and objective quality evaluation parameters both are greatly improved,and effective information such as spectral information and edge feature information is well preserved.