Research On Phase Calibration Method Of Multi-baseline Tomography SAR

Due to the interference of air flow and other factors,the radar sensor in the multi-baseline airborne SAR system will inevitably generate a position shift during flights.The motion error causcd by this position shift will affect the position of the radar antenna phase center and the direction of the radar beam,resulting in the phase error of the SAR data,thus affecting,the quality of the SAR tomography.In order to ensure the quality of SAR tomography,it is necessary to calibrate the phase error caused by the sensor motion deviation in the SAR data processing step.The thesis takes tomography of multi-baseline SAR data as the main line,and focuses on the phase calibration method in the tomography process.Combining simulation data and ESA BioSAR 2008 multi-baseline data for verification experiments,the specific research content and results as follows:(1)The phase calibration method based on the Permanent Scatterer(PS)technology is analysed.Using the advantage that the scattering characteristics of the permanent scatterer will not change within a certain period of time,combined with the basic principles of PS-lnSAR,thc interference phase error model of the permanent scatterer is analyzed.Only the phase error of permanent scatterers is required,and then the phase error of the whole image is obtained by interpolation and then realize the phase calibration.The results of spectrum tomography show that this method can eliminate the phase error to some extent,but the side lobes are relatively large,and the overall effect is not ideal.(2)The phase calibration method based on Phase Center Double Location(PCDL)technology is analyzed.Based on the signal model of SAR,the position error of the phase center of the sensor antenna in the line of sight is analyzed,and the mathematical expression of the phase error related to the position error is established.Combining with the phase linking algorithm,the iterative process of the double location method of the phase center is deduced.The phase error of the whole image is calculated by the iterative formula,and then the phase calibration is realized.It is found that the side lobes in the spectrum tomograms are relatively less,the amount of spectrum information in the canopy is relatively more,and the overall effect is better.(3)The weight calculation model is improved based on the PCDL algorithm.The Euclidean distance between pixels is used to determine the similarity,and then the weight calculation is improved.The effectiveness of this method has been confirmed through experiments.Not only are there relatively few side lobes in the spectrum tomography,but also the amount of spectrum information in the canopy remains relatively complete,and the overall effect is the best.