The Real-time Processing Of Multi-mode SAR&GMTI Imaging Based On TMS320C6678

Synthetic Aperture Radar(SAR)can realize high-resolution and long-distance imaging of ground stationary targets in any weather condition and any time period,and can also detect ground moving targets.Therefore,the airborne synthetic aperture radar has been widely used in the military and civilian fields.In this thesis,a real-time imaging processing system based on TMS320C6678 is built to achieve real-time imaging of multiple SAR modes and targets for GMTI(ground moving target detection).Real-time imaging of strips SAR and round spotlight SAR and GMTI of real-time processing functions are realized.At the same time,for the problem of SAR image storage and transmission,An improved DCT transform image compression real-time algorithm based on TMS320C6678 is implemented.The measured data processing experiment effectively validates the real-time imaging processing system and imaging algorithm.First of all,from the research background and development status of SAR imaging,this thesis briefly introduces SAR imaging algorithm,moving target detection algorithm and radar processor.The radar processor is a hardware processing platform for synthetic aperture radar imaging and is used to complete the digital signal processing tasks of the radar system.The self-developed radar processing platform is a digital signal processing system based on the 3U and VPX bus architecture.The system can implement data acquisition functions,data storage functions,signal processing functions and data playback functions.The system consists of a hardware combination and embedded software processing unit.The hardware combination includes multiple function modules.The embedded processing software mainly implements radar data processing.Data processing is mainly completed by a multi-core processor chip on TMS320C6678.The chip adopts a new Harvard architecture and supports fixed-point and floating-point operations.At the same time,the chip integrates eight independent processing cores,which supports a maximum frequency of 1.25 GHz.In addition,it has a rich external interface and a three-tier storage architecture,which improves data communication and resource sharing between devices.Secondly,this thesis mainly introduces the implementation of strip SAR,bunching SAR imaging algorithm and moving target detection processing on DSP.Among them,strip SAR imaging processing includes inertial parameter calculation,motion compensation,azimuth tone frequency estimation,range-to-range migration correction,azimuth pulse compression,image quantization multi-view,and image interpolation.The spotlight SAR algorithm consists of several steps including inertial parameter calculation,motion compensation,distance FFT,Keystone transform,distance IFFT,azimuth tone frequency estimation and azimuth pulse compression.GMTI processing can be achieved through such steps as motion compensation,azimuth FFT,CFAR detection,clustering,sidelobe masking,and parameter estimation.Finally,based on the limitations of data transmission bandwidth,direct transmission of SAR images can not meet the real-time requirements of SAR systems.Therefore,this thesis performs JPEG image compression processing on the SAR imaging results,which focuses on an improved DCT transform method.The improved method is to cyclically expand a large number of loop nested operations,and use the table lookup index method to achieve the purpose of optimizing the DCT transformation time.Finally,the measured data prove that this improved method can reduce the time for JPEG image compression.