Implementing Synthetic Aperture Radar Algorithm Algorithm Based On Domestic Processor

Synthetic Aperture Radar(SAR)imaging has the characteristics of all-weather,all-weather,long-distance and high-resolution imaging,which can greatly improve the radar's information acquisition capabilities.Airborne SAR imaging systems using aircraft as carriers have been widely used in military and civilian fields.SAR imaging processes a large amount of data,complex calculations,and high requirements for processing system performance.And with the development of domestic processors and the demand for localized replacement,the research on real-time processing of airborne SAR imaging algorithms in domestic processors is of great practical significance.First of all,the principle of airborne SAR imaging and its improvement methods are discussed.Firstly,the geometric model and signal model of airborne SAR imaging are summarized.Then two distance compression algorithms are introduced: matched filtering method and the principle and realization method of dechirp method under high resolution and large bandwidth signal.The thesis analyzes the principle of azimuth focusing through the Back Projection(BP)algorithm,and then proposes an improvement method.Finally,for the motion error of the platform,the motion compensation method based on the inertial navigation equipment information and the phase gradient autofocus(PGA)algorithm is introduced,and a three-step motion compensation method is proposed.In order to improve the image perception after SAR imaging and solve the problems of the image's observation domain,brightness,resolution and continuity,the thesis analyzes the algorithm principles and implementation methods of geometric correction,brightness equalization,multi-view quantization and image stitching.Subsequently,the thesis introduces in detail the implementation of airborne SAR imaging on the domestic processor platform.First,the topological structure of the nationally produced hardware platform and the resources of domestically produced digital signal processors(DSP)and domestically produced field programmable gate arrays(FPGA)are introduced.Then the DSP algorithm implementation plan is designed in detail,including multi-DSP parallel processing flow control,DSP multi-core parallel processing control method,fast data access method through cache and DMA,fast matrix realization method and joint processing through FFT accelerator and core Fast FFT implementation method,performance comparison and effect analysis are given.Next,the FPGA system control implementation plan is proposed,including DSP power-on sequence control,DSP initialization configuration control,original data packet forwarding function,multi-chip processor high-speed data exchange and other functions.Finally,the entire system is verified through experiments with measured data,and the effects of different algorithm steps and their effects on imaging results are compared and analyzed.Firstly,compare the effects of motion compensation based on inertial navigation information and PGA to verify the effectiveness of the three-step motion compensation method;Then compare the effects of direct interpolation and interpolation through chirp-z transform(CZT),and get the effectiveness of quickly achieving Range Cell Migration Correction(RCMC)through CZT;then through the perception analysis of the measured images,the effects of geometric correction,brightness balance,multi-view quantization and image stitching algorithms are verified.Finally,a complete imaging rendering of the whole system is given.