Research On Registration And Orthorectification For Space-borne InSAR Image

Interferometric Synthetic Aperture Radar(InSAR)technology obtains high-precision digital elevation model and derivative data products by extracting the phase information of the image.It has all-time,all-weather,and large-area features,and is widely used in many fields as terrain mapping,military reconnaissance,and disaster monitoring.It occupys an important position in the fields of national economic construction and national security.In this paper,we focus on InSAR data processing methods for satellites,and study the two aspects of InSAR complex image registration and SAR orthorectification.The research content mainly includes:(1)Based on the analysis of three common registration measure functions for InSAR complex image registration,an adaptive registration algorithm model based on DEM(Digital Elevation Model)is constructed.The algorithm can flexibly select registration by calculating the slope information of the registration area.The measure function solves the problem of low accuracy of registration due to the single selection of the measure function under complex terrain.Experimental results show that the proposed algorithm performs better than other algorithms.(2)The RD(Range Doppler Model)positioning model is proposed based on the positive solution orthorectification method.This algorithm which combines the forward solution orthorectification algorithm with the indirect orthorectification algorithm.The correction results show that in addition to eliminating the “hole” pixel,it also effectively compensates the geometry of the pre-corrected image,such as overlaid shadows and shadows.The corrected image is in good agreement with the Google image in the same area and achieves an ideal orthorectification effect.(3)Using VS2010 as a development platform,combining the C++ language to implement the algorithms of Chapters 3 and 4.usingA large number of measured data verify the accuracy of the registration algorithm and the orthorectification algorithm.The results show that the DEM-based adaptive registration algorithm obtains a registration image with an azimuth accuracy of 0.0269 pixels and a distance accuracy of 0.0398 pixels,which fully satisfies the image interferometric requirements.The accuracy of the corrected image obtained by the indirect orthorectification algorithm based on the RD localization model is about 0.8 pixels,and the correction result is good.