The Processing Of SAR/InSAR High-speed Non-stationary And Real-time Imaging

The synthetic aperture radar(SAR)with a good working performance in all weather and all time has an especially unique impact of the military field.With the extension of the application,it also plays a more and more important role in the development of national economy.Interferometric synthetic aperture radar is a new radar imaging technology which is based on synthetic aperture radar.InSAR gets the digital elevation map by processing two different-perspective focused SAR images,and it is mainly used for terrain mapping and surface deformation monitoring and other fields.The processing of InSAR is complex and computationally intensive,which makes the InSAR real-time processing system have high requirement for both software and hardware platforms.In this paper which is based on the InSAR signal real-time processing system,combine the advantage of FPGA in parallel processing and the strong performance of TMS320C6678 multi-core DSP in the floating point operation,and finally realize the real-time processing of InSAR.This paper mainly study the radar echo signal imaging and interference processing,when the platform is in the high-speed and non-stationary state of motion.The two-dimensional spectral expression of the signal is obtained by using the series inversion method,and the two-dimensional solution of the azimuth and the distance are achieved.Then,using the high-order phase filter processing to achieve the azimuth frequency offset compensation and the realization of the signal focus in the direction.In the interference algorithm,for InSAR operating mode with short baseline,the traditional InSAR processing flow is optimized,since the platform is very large from the target and the baseline is very short,the step of image registration will be ignored.This can effectively reduce the amount of computation in real-time processing.In this paper,the optimized InSAR processing flow consist of a total of 6 processing steps and finally obtain the digital elevation diagram.The 6 steps are the interference phase generation,mean filtering,mass map generation,phase unwrapping,phase elevation,grid interpolation.In the realization of the algorithm,FPGA is responsible for completing the scene data interception,parameter processing,part of the data operation and the final data transmission.In the DSP processing,by taking full advantage of TMS320C6678 multi-core advantage,the rational allocation of tasks on each core,and inter-processor communication for multi-core collaborative work,we achieve the maximum realization of multi-core DSP parallel processing.In the implementation of the task,each core fully utilizes the inline function and the DSP computing unit to ues the software pipeline to a great extent.And finally realize the real-time imaging and interference processing of the signal.