Research On Algorithms Of SAR And Optical Image Fusion

With the rapid development of multi-source remote sensing technology,the application of Multi-source image fusion has increased.SAR and optical images have their own characteristics,quite different from each other.SAR images can break out the constraint of visible optical images and detect more imperceptible feature information.Optical images are appropriate used for physical geography.However,due to the limitations of sensor object characteristics,unable to obtain radar satellite images with spectral information.Therefore,how to organically combine different images,to achieve the complementary advantages of image by the technology of image fusion has become an important research topic in the development of image applications.This thesis embarks from the imaging mechanism and fusion significance of SAR and optical image,has analyzed of the SAR and optical image fusion algorithm,and in-depth study of SAR and visible image fusion.The main work is summarized as the following three aspects:(1)The existing fusion algorithms are summarized,and the advantages and disadvantages of the existing algorithms are analyzed from both subjective and objective aspects.(2)In order to avoid the excessive extraction of SAR image feature information and reduce the spectral distortion.In this thesis,aims at the problem of excessive extraction in SAR image feature extraction based on gray level gradient algorithm.A fusion method combine adaptive HIS with phase congruency method is proposed.The phase congruency algorithm can distinguish the target contour and texture edge using phase information,it can effectively avoid the excessive extraction of the edge,and the adaptive HIS algorithm can overcome two image gray value differences,to achieve a smooth transition to the information,reduce the spectral distortion of the image fusion.(3)In order to reduce the blurring of the edges in the fused image which caused by adaptive HIS with phase congruency method,a fast fusion algorithm based on the original image similarity is proposed.The algorithm obtained the fusion image by establishing the statistical relationship between the ideal fusion image and the original SAR and the visible image.Results are illustrated the spectral and spatial character of fused image are better balanced.A dataset fusion of Landsat-8 optical image with Sentinel-1A SAR image to demonstrate that the proposed method has good performance by compared with the existing algorithms.Experimental result shows that the proposed method can effectively inject the missing features into optical image as well as better to maintain spectral and spatial structure quality,achieve a better visual effect,and improve the capability of target detection.