Study On Fully Polarimetric SAR/InSAR Data Calibration Technology

Fully polarimetric synthetic aperture radar(SAR)system can acquire the fully polarimetric scattering information of terrain by transmitting orthogonal electromagnetic waves.On the foundation of above,polarimetric interferometric SAR(InSAR)system can further acquire the interferometric information.Compared with the single polarimetric SAR system,the capacity of fully polarimetric SAR/InSAR system has been greatly enhanced in acquiring terrain target characteristics.So far,fully polarimetric SAR/InSAR technology has been widely used in agriculture,forestry,ocean,disaster reduction and traffic etc.The acquired fully polarimetric SAR/InSAR data always has errors of channel crosstalks,amplitude and phase imbalances.Therefore,fully polarimetric SAR/InSAR data must be calibrated before used,which results in the requirements for polarimetric SAR/InSAR data calibration.This dissertation is dedicated to make a research on the existing problems in polarimetric SAR/InSAR data calibration.The main contributions are as follows.(1)In Chapter 3,a research is made on the problem that Ainsworth algorithm is sensitive to channel noise.The reason for the problem is analyzed in theory and an improved Ainsworth algorithm is presented,which is insensitive to the channel noise.On the basis of considering different cross-polarization channel noise powers and low signal noise ratio,the improved algorithm modifies the estimation formula of the cross-polarization imbalance parameter and makes a revision to original iteration process.The experimental results show that the improved algorithm has a good parameter estimation performance.Combined with the advantages that Ainsworth algorithm has less assumptions and can keep the polarimetric orientation angle,the improved algorithm can be a priority chosen in case of low crosstalk.(2)For the polarimetric calibration problem under high crosstalk situation,the following works have been done in Chapter 4.The Capabilities of Quegan,A and Az algorithms are analyzed in theory and the variations of their performances with crosstalks are given using simulated data.Based on the results from theatrical and simulated analyses,the following conclusions can be conducted:in case of high crosstalk,the Az algorithm has the best performance,Quegan algorithm second and A algorithm worst.A non-iterative polarimetric calibration algorithm for high crosstalks is presented.On the foundation of keeping the higher order terms of crosstalk parameters,the algorithm establishes a system of non-linear equations about crosstalks.By solving the equations,the distortion parameters can be obtained accurately and directly.In order to further solve parameter estimation problem under both low signal noise ratio(SNR)and high crosstalk,a channel noise estimation algorithm is introduced.The experimental results show that the non-iterative polarimetric calibration algorithm considering channel noise estimation can realize the accurate estimation of polarimetric distortion parameters and channel noise,even when the crosstalks are high and the SNR is low.(3)For the co-polarization channel imbalance estimation problem when lacking corner reflectors in observing scene,the following studies are conducted in Chapter 5.The polarimetric scattering characteristics of bare soil under high frequencies like X-and Ka-bands are analyzed theatrically and then verified by using the real X-band airborne data.Based on the theatrical and simulated analyses,the following conclusions are obtained:the HH amplitude and phase differences between HH and VV channels are very small for bare soil and the helix component in backscattering energy is very low.Therefore,the bare soil can be chosen as an effective distributed calibrator in calibration process.A co-polarization imbalance determination algorithm for high-frequency data is presented.The algorithm reduces the two-dimensional search in Shi algorithm into one-dimensional search by adequately using the phase characteristics and thereby can realize the stable estimation of co-polarization imbalance.Meanwhile,the algorithm limits the solver range of co-polarization imbalance by using the amplitude characteristics of bare soil,which further enhances the robustness of the algorithm.The experimental results with real airborne data validate the proposed algorithm.(4)For the covariance matrix estimation and polarimetric InSAR phase inconsistency problems existing in polarimetric SAR/InSAR data calibration processing,the following works are done in Chapter 6.A calibration pixels extraction method is proposed,which combines the information of the pixel’s power and correlation between co-and cross-polarizations.Firstly,the low power pixels are removed by using the power information.Then,the pixels that used in covariance matrix estimation are selected through the correlation between co-and cross-polarizations.Finally,the high power pixels are excluded in order to eliminate the influence of saturated pixels on covariance matrix estimation.The experimental results based on real data show that the application of the proposed method in fully polarimetric data calibration process can effectively improve the parameter estimation stability and processing capability for different scene data of the calibration algorithms.A polarimetric interferometric phase correction algorithm is presented.The algorithm solves the cross-polarization channel interferometric phase inconsistency problem caused by parameter k using the statistical characteristics of polarimetric differential interferometric phase.Furthermore,the polarimetric interferometric phase inconsistency problem arising from the residual phase error in polarimetric calibration is solved by using the characteristics that the polarimetric differential interferometric phases are zero in areas that have no vertical structure distribution.The experimental results with real data show that the proposed algorithm can effectively correct the polarimetric interferometric phase inconsistency problems.