RESEARCH ON SAR IMAGE MATCHING TECHNOLOGY BASED ON SIFT

摘要

Image matching is one of the key technologies in remote sensing image fusion and navigation. Image matching of Synthetic Aperture Radar (SAR) is a process to find relationship of pixels in many SAR images, which directly involves and affects the application of SAR image in many areas such as mapping, integrated navigation and image fusion. In order to improve the searching speed in matching, pyramid strategy is used. Considering the characteristic of low S/N ratio in SAR image, curvelet is introduced in preprocessing SAR images. 1,For the great metamorphoses between SAR images, direct Scale Invariant Feature Transform (SIFT) is used in matching of destination image and referenced image which are similar with each other in greyscale. 2, edge extraction is implemented in SAR images acquired at different times and on different orbits by using Canny operator, and then SIFT key points is extracted to match the images. Combined with correlation coefficient controlling method, error matching points are wiped off and good result is acquired.