Research On Airborne SAR Image Block Adjustment

Synthetic Aperture Radar(SAR)with all-day,all-weather imaging capabilities,widely used in cloudy,foggy and rainy areas for rapid terrain mapping.Block adjustment is a key step to realize terrain mapping of a large area with sparse ground control points(GCPs)using SAR images.In order to improve the practicalization of airborne SAR image block adjustment,this paper focuses on tie points(TPs)extraction for SAR images with same side-looking direction,airborne SAR image block adjustment based on trajectory constraint equations,airborne InSAR(Interferometric Synthetic Aperture Radar)block adjustment with weights determined based on posterior variance component estimation,and conducts related experiments using spaceborne SAR images,airborne SAR images and unmanned aerial vehicle(UAV)SAR images.The main work and innovations of the thesis are:1.The background and significance of SAR image block adjustment are analyzed.The research status of related technologies is combed.The main problems for the practicalization of airborne SAR image block adjustment are analyzed.2.A TPs extraction method for SAR images with same side-looking direction from different trajectories based on differential constraints is presented.Image pyramids are built for two SAR images,and then these images are layer by layer matched by computing correlation coefficient using rectangularly matching window with increased azimuth side.False matching points are removed by RANSAC algorithm with differential constraints.Coordinates for points in lower pyramid images are estimated based on global bilinear transformation model for coordinates in azimuth and local bilinear transformation model for coordinates in range.Experiments performed on spaceborne ENVISAT ASAR images,Chinese airborne SAR images and Chinese UAV SAR images confirm that the proposed method can obtain matching points(TPs)with uniform distribution and high correct rate.3.An airborne SAR image block adjustment method based on trajectory constraint equations and Schreiber Rule is proposed.Trajectory constraint equations are established using TPs in overlapping regions of SAR images from same trajectories,and equivalent error equations for TPs are obtained based on Schreiber rule.Experiments performed on Chinese UAV SAR images and Chinese airborne SAR images show that the proposed method can effectively improve the accuracy of the adjustment on range and elevation direction with slower convergence speed,and the conclusions are validated by experiments on varying the distribution and number of GCPs.In addition,the construction of equivalent error equations can help decrease the dimensions of the coefficient matrix in normal equations and improve the computational efficiency of the adjustment.4.Airborne InSAR block adjustment methods based on the extended model of variance components estimation of Helmert type for weights determination of original error equations and Schreiber Rule are proposed respectively.Aiming at the problem that the variance component for each observation component cannot be estimated directly,weights for original error equations are determined based on the idea of posterior variance component estimation.Equivalent error equations for TPs are obtained based on Schreiber rule.Experiments performed on Chinese dual antenna airborne InSAR data show that the method with posterior variance component estimated can take generation accuracy and mosaic accuracy of DEM into account reasonably,and the construction of equivalent error equations can help decrease the dimensions of the coefficient matrix in normal equations and improve the computational efficiency of the adjustment.