Research On Multidimensional Imaging Technology Of Spaceborne Bistatic SAR

Differential Synthetic Aperture Radar Tomography(D-TomoSAR)is a new technology for space observation in recent years.It can be used to acquire the elevation and deformation information of ground surface by high-precision multidimensional imaging.At present,this technology has been widely used in geological hazard warning,geophysical parameter inversion and urban planning.At the same time,spaceborne SAR system is developing towards multi-static,multi-mode,and high spatial resolution,which provides a great opportunity for the further research of D-TomoSAR.In this paper,the spaceborne bistatic SAR(BiSAR)is taken as the background,the key and difficult problems of SAR multi-dimensional imaging by using D-TomoSAR are researched.Some solutions and processing algorithms are proposed in the aspects of high-resolution two-dimensional(2-D)imaging of spaceborne BiSAR,estimation of deformation parameters under different deformation conditions,and retrieval of 3-D deformation parameters.Those researches provide a theoretical basis for promoting multi-static SAR to realize multidimensional imaging and promoting the further practical application of D-TomoSAR.The main work and innovations of this paper are as follows:1.Aiming at the problem of high-resolution 2-D imaging of spaceborne BiSAR in long synthetic aperture time,a new high-order range model based on the relative motion state of SAR platforms and ground target is proposed to accurately describe the slant range history between the platform and ground target.And on this basis,an extended azimuth nonlinear chirp scaling(EANLCS)algorithm with an addition of highly varying residual Doppler centroid correction is proposed to realize highresolution 2-D imaging of spaceborne BiSAR by analyzing the 2-D spatial variance characteristics of the echo signal,which provides a foundation for the multidimensional imaging of spaceborne BiSAR.2.D-TomoSAR processing with high-resolution X-band SAR data provides even higher accuracy for monitoring of deformations of the surface.However,the use of higher frequency data increases the sensitivity of the system to the ground minute deformation,resulting in mismatching between the actual deformation situation and the assumed linear deformation model,which cause a large estimation error of deformation parameters.Therefore,this paper first studies the acceleration/deceleration deformation trend of the ground,which is one of the forms of the minute deformation,and an extended D-TomoSAR signal model is proposed.By analyzing the characteristics of the proposed signal model,the process of solving the deformation parameters is transformed into a problem of estimating the coefficients of 2-D polynomial phase signals(2-D PPS).Moreover,considering the unique nonuniform and undersampling characteristic of D-TomoSAR signal,the RELAX algorithm is introduced into the traditional 2-D product high-order ambiguity function(PHAF)algorithm to estimate the ground elevation and the acceleration/deceleration trend of deformation.Compared with the traditional method based on compressive sensing(CS)technology,the proposed algorithm has less computational complexity.3.In addition to the acceleration/deceleration deformation trend of the surface,the seasonal deformation caused by the thermal dilation of structures is also an important component of the minute deformation,which cannot be neglected.Therefore,on the basis of the above research,a generalized D-TomoSAR signal model is proposed to estimate the seasonal deformation of scatterers.Then,the process of solving the deformation parameters can be transformed into a problem of estimating the parameters of a 2-D hybrid sinusoidal frequency modulated and polynomial phase signal(2-D HSFM-PPS)by analyzing the characteristics of the proposed signal model.Subsequently,inspired by the quasi maximum likelihood(QML)principle,an effective algorithm for estimating the parameters of 2-D HSFMPPS is proposed based on the sparse time-frequency reconstruction.Compared with the traditional CS-based algorithm,the proposed algorithm has lower signal noise ratio(SNR)threshold and less computational complexity.4.Considering that the traditional D-TomoSAR technology can only obtain the deformation parameters of the target along the line of sight(LOS)of radar,thus,the real deformation of the surface can not be accurately described.Therefore,the problem of retrieving 3-D surface deformation parameters(East-West,North-South,Up-Down)is also studied based on the above researches.Aiming at the inaccuracy of the estimation of North-South motion component using the traditional methods,an improved D-TomoSAR signal model is proposed by introducing the imaging squint angle of spaceborne BiSAR into the traditional D-TomoSAR processing.As a result,the relationship between 3-D deformation parameters and the phase of SAR single looking complex(SLC)is established.Then,the 2-D PPS phase coefficient estimation algorithm is proposed to estimate the 3-D deformation with equal-precision by analyzing the proposed signal model.Compared with the traditional methods,the proposed algorithm reduces the amount of data required,and greatly improves the estimation accuracy of North-South deformation component without changing the current satellite orbit,which provides a new solution for retrieving the 3-D deformation parameters with equal accuracy.