Synthetic Aperture Radar Real-time Imaging Algorithm Optimization And System Development

The increasingly sophisticated modern battlefield environment brings even more serious challenges to radar precision guidance.In order to accurately extract targets from a complex battlefield environment,synthetic aperture radar must be used to obtain high-resolution images of suspected targets and target recognition must precede from SAR images,thus,accurate punching can be realized.On the one hand,due to the strong maneuverability of the missile-borne platform and the inability to be equipped with high-precision inertial navigation system,there will be a large number of motion errors in the echo data,which results in a high complexity of the missile-borne SAR imaging algorithm.On the other hand,the distinctiveness of missile-borne application requires Imaging algorithms must be processed in an extremely short period of time.In this case,real-time imaging of missileborne SAR faces enormous challenges.In order to overcome the bottleneck of real-time imaging of missile-borne SAR,this paper studies the real-time imaging algorithm optimization and system development of missile-borne SAR.Firstly,this paper studies the missile-borne SAR imaging algorithm and the principle of inertial data motion compensation algorithm,which lays a theoretical foundation for the subsequent algorithm implementation.Secondly,this paper studies the architecture of TMS320C6678 multi-core DSP chip,two programming methods of bare metal and BIOS operating system and optimization of bare-metal implementation have been proposed.The algorithm of missile-borne SAR imaging is implemented based on multi-core DSP platform,and a set of missile-borne SAR signal processing system is constructed.Next,in order to further improve the real-time performance of the missile SAR imaging algorithm,this paper studies the hardware architecture and programming method of the embedded GPU,and implements and optimizes the missile SAR imaging algorithm based on the embedded GPU platform,and compares it with the multi-core DSP implementation.Nearly 9 times speedup has accomplished,which shows that the GPU has obvious advantages in computing acceleration.In order to promote the application of embedded GPU in the missile platform,this paper designs the hardware and software architecture of the missile SAR signal processing system based on embedded GPU.Finally,in order to verify the implementation of the missile-based SAR imaging algorithm based on multi-core DSP,a set of digital simulation test system software is designed and developed.The software system is implemented by Qt GUI library and forms a semiphysical simulation test system together with the missile-borne SAR signal processing system.The running and testing results of the semi-physical simulation system show that the real-time imaging algorithm of the missile-borne SAR described in this paper produces good imaging quality and meets the real-time requirements.