Realization Of SAR Real-time Processing Technology Based On FPGA

Synthetic aperture radar(SAR)as a high-resolution remote observation method.,it also has the characteristics of all-weather and all-weather.It is widely used in military fields such as target reconnaissance,battlefield situation perception as well as civil fields such as disaster detection,environmental monitoring and so on.In these applications area,the real-time and miniaturization of SAR imaging has gradually become a hot issue of research.With the development of semiconductor technology,digital signal processing devices such as FPGAs,DSPs,and ASICs are used to accelerate signal processing.FPGAs are widely used in SAR imaging processing because of their high density,parallel processing architecture,and low power consumption.This article mainly researched on the FPGA implementation of SAR real-time imaging processing.This paper first introduces the background and significance of SAR imaging,the development and application of imaging algorithms and FPGA devices are described in detail.At the same time,it outlines the structure of the article.Next,the data acquisition model of spotlight SAR is introduced.Combining with the data acquisition model,the imaging processing steps of polar coordinate format algorithm(PFA)are analyzed.The range data is resampled with PCS principle,and the azimuth data is resampled by the SINC interpolation.Because PFA assumes that the radar platform is moving at a uniform speed,in actual conditions,due to the influence of air flow disturbances and non-uniformity of electromagnetic wave propagation media,the focusing effect is difficult to meet the requirements.In order to improve the quality of focus,the phase gradient autofocus(PGA)algorithm is added.In the case of long synthetic aperture or large angle of view,the performance of the PGA to estimate the full-aperture phase error will be reduced.Therefore,the method of molecular aperture is used and the sub-aperture offset(Map Drift,MD)algorithm is used to improve the subaperture phase error accuracy.Aiming at the problem of migration of residual distances across multiple distance elements,a two-dimensional autofocus algorithm under polar coordinate format was studied.Two-dimensional phase error was obtained by using the estimated azimuth phase error.Finally,a block-wise PGA method in image domain is introduced to reduce the impact of space-variant phase error.In order to accelerate the process of imaging and self-focusing,we build a parallel system architecture based on Ethernet and DDR3 on FPGA.Ethernet implement the transmission of data between board and the host computer,the signal processing module mainly includes the range data resampling module,the azimuth data interpolation module,the PGA submodule and the cross-correlation submodule.DDR3 completes the storage of a large number of SAR data.To confirm validity of the design,a verification system based on Ethernet,FPGA and DDR3 SDRAM.The architecture of the parallel processing system is introduced,and the main state machine processed by this system is given according to the flow of data.The imaging processing results,data processing accuracy and imaging processing speed of the system are verified on the processing board centered on the Virtex7-XC7VX690 T.The measured data shows that the system has good imaging results,and the processing accuracy and speed meet the requirements.Finally,it summarizes the full-text work and points out the future research direction.