Design And Development Of SAR Time Domain Imaging Algorithm

The development of synthetic aperture radar imaging technology is promoted by the demands in civil and military fields.The research background of this thesis is the application of synthetic aperture radar imaging algorithm in the missile platform.Under the missile platform,the SAR imaging algorithm needs to work in high-squint which leads to severe distance and dimensional coupling of radar echo signal.In the case,the focusing performance of SAR imaging algorithm is reduced.The time-domain imaging algorithm has the advantages of flexible observation geometry,high focusing accuracy and controllable image resolution which can effectively guarantee the focusing accuracy of SAR imaging.However,the practical application of time-domain imaging algorithm is limited by its huge computation.Therefore,it is of great significance to study the design and development of time-domain imaging algorithm.In this thesis,the design and development process of time-domain imaging algorithm is described in detail from three aspects.First,the principle of time-domain imaging algorithm.Second,the design of algorithm implementation system.Third,algorithm hardware implementation.The principle of accelerated fast factorized Back-Projection imaging algorithm based on spectrum fusion is firstly introduced and the implementation flow of the algorithm is given.In order to solve the problem of motion error of missile platform,the motion compensation algorithm based on inertial navigation data and Map drift algorithm based on echo data are described in detail.In the aspect of algorithm engineering,considering the high parallelism of AFBP time domain algorithm itself,this paper designs a parallel architecture based on three DSPs.According to the time-consuming of each imaging module in the imaging algorithm,the workflow and the direction of data flow of the algorithm system are determined.The data transmission between three DSPs is completed through the use of three high-speed interfaces like SRIO,Hyperlink and PCIe.At last,it discusses the design of communication mode between cores which lays the foundation for the final engineering implementation of the algorithm.In the aspect of algorithm hardware implementation,firstly,the memory of DSP is managed through the CMD file.According to the process of AFBP imaging algorithm,the parallel data processing model based on data flow is selected.The processing tasks of each DSP chip are allocated reasonably according to the parallel processing architecture.Then,the real-time performance of the algorithm is improved by using the code optimization methods such as linear assembly function,look-up table method and compiler optimization Implementation efficiency.Finally,through the analysis of DSP imaging results,the correctness of AFBP time domain imaging algorithm design and development based on multi-core DSP is verified.