Study On Very-High-Resolution Of Airborne/Spaceborne SAR Imaging

Very-high-resolution synthetic aperture radar(VHR-SAR)can detect microstructure information of target and obtain the fine characterization.It has unique advantages in the research of VHR-SAR imaging,multiple views information fusion,moving target detection,target recognition.This paper research on VHR-SAR imaging of the airborne and spaceborne platforms.In airborne VHR-SAR,errors of several subbands need to be estimated accurately to achieve the whole bandwidth in the range;motion error need to estimated to achieve very-high resolution SAR imaging in azimuth.Airborne VHR-SAR plays an important role in high-accuracy battlefield reconnaissance.Due to the high orbit,spaceborne VHR-SAR has become an important remote sensing technology due to outstanding performance of global observation in day and night under all weather.For the development requirements,this paper studies subbands synthesis in range,multi-mode VHR-SAR imaging mounted on airborne and spaceborne platforms,VHR-SAR imaging,target detection and fine imaging under unstable platforms.Specifically,the main contents of this paper are as follows:1.For bandwidth synthesis,a processing flow of bandwidth synthesis based on error estimation is introduced in the mode of overlapped-bandwidth.The phase errors are estimated and compensated using the overlapped subbands and the influence of constant phase is discussed in detail.For non-overlapped-bandwidth,the side-lobe balanced model is proposed to estimate phase before bandwidth synthesis.First,to achieve error compensation of a subband,a range phase adjustment by contrast enhancement algorithm is proposed to estimate inner-channel high-order phase errors.Second,a sidelobe balanced model is established to deduce the relationship between the balanced sidelobe and constant phase error.so the constant phase error can be directly obtained in an efficient manner to compensate the low-order phase errors.Experimental analysis using real data demonstrates the effectiveness of the method.2.For airborne multi-mode VHR-SAR,high-precision INS is applied to estimate the motion error and the space-variant phase errors are obtained by interpolating fitting.For spotlight VHR-SAR,a modified RAM algorithm is proposed,the modified Stolt interpolation kernel can be well combined with motion estimate algorithm.A method is proposed to estimate phase errors based on sub-aperture error estimation and full aperture error fitting.At the same time,a flowchart of VHR-SAR imaging combined with error compensation is given.For high squint spotlight VHR-SAR imaging,PFA algorithm based on polar coordinates interpolation is given.At last,for the fusion of multi-view SAR,a new two-steps matching method is proposed,in the first step,the track information is used to obtain the difference of the multiple perspectives,and then used to achieve coarse registration.Secondly,a modified SIFT algorithm is used to detect,depict,and match the key points of SAR images.Finally,the multiple perspectives fusion is achieved.3.For spaceborne multi-mode VHR-SAR,a modiefied BAQ decoding algorithm is proposed which is based on GPU framework.The slant distance and equivalent velocity can be computed precisely with geometric configuration and orbit parameters.For dual-channel SAR,this chapter proposes a "coarse to fine" channel error correction method.the channel correction method is introduced to achieve amplitude and phase error correction,the STAP is used to correct the parts of channels errors,and the function called azimuth ambiguity suppression ratio is proposed to estimate residual channel errors.A modified PFA algorithm based on wavefront curvature correction is proposed.space-variant geometric distortion and defocusing caused by wavefront curve are compensated by interpolation and filtering,respectively,which improves the imaging area.4.For high-resolution SAR imaging of non-stationary platforms,the vehicle-borne microwave photonic radar is introduced here,which is vibrational caused by the unevenness of the road surface,resulting in a non-periodic envelope deviation error.The conventional motion compensation method cannot compensate effectively.This part proposes a point tracking envelope detection and compensation method,which can eliminate the non-periodic high-frequency errors and achieve foucs images.Aiming at the problem of Helicopter SAR azimuth ghosting,the expression of SAR echo based on helicopter vibration platform is deduced,Jacobian-Angle constant equation is introduced and the first order Bessel series expansion is performed.Azimuthal ghosting is present when imaging is caused by vibration.The error estimation method based on echo data is proposed.The frequency points of the vibration are extracted and the corresponding error phase is inverted.After compensation,the periodic high-frequency error can be eliminated and the azimuth without blur imaging can be achieved.5.For SAR target detection and fine imaging,a two-step “coarse-to-fine” target detection method is proposed for ship detection with spaceborne VHR-SAR images.In coarse detection,pre-processing and dual-parameter CFAR algorithm is performed,fine detection was performed based on APACE algorithm,the difference in entropy values can be used to achieve fine detection.Then,a super-resolution imaging method is proposed to solve the problem which extracting targets and enhancing resolution under strong clutter in SAR image.Based on singal correlation,an adaptive clutter threshold is proposed to extract strong scatterers of targets and weak scatterers of targets are extracted by gradient descent method.Finally,target resolution is enhanced to realize super-resolution imaging by simplified regularization method.Real-measured of data reflectors and stationary vehicle validate the effectiveness of the proposed method.