SAR成像实时处理平台的设计与实现

发布时间：2018-03-04 13:46

本文选题：雷达实时成像　切入点：TMS320C6678　出处：《西安电子科技大学》2014年硕士论文　论文类型：学位论文

【摘要】：本文结合雷达成像的研究背景,分析了成像处理系统的实际需求和新发展,提出并实现了一种新型合成孔径雷达成像平台的设计,用以满足合成孔径雷达实时成像的实时性、高分辨率和大数据量要求。本设计采用了两片DSP-TMS320C6678+FPGA+VPX架构作为雷达成像处理平台,与其他功能模块的板卡通过VPX标准接口互联,共同完成雷达系统的成像功能。成像处理平台的DSP是TI公司新推出的高性能八核DSP-TMS320C6678,单核主频高达1.25GHz,支持定点和浮点高速运算,具有三个独立的DMA控制器,和丰富的内存资源、外部存储扩展接口和多通道高速串行接口。在硬件设计方面,着重介绍了电源和时钟模块的设计,在满足雷达成像平台的功耗和时钟需求前提下,减少板卡的功耗和体积。接下来基于雷达成像处理平台,对板卡上的高速接口进行了详细分析。本板卡的高速接口包括DDR3、Hyperlink、PCIe、SRIO和千兆以太网。首先介绍了高速接口的基本原理和互联设计,再具体的分析了接口通信软件的配置和调试,最后对本设计的高速接口进行速度测试和分析。此外,还介绍了板卡的普通接口如SPI和GPIO,并提出和实现了其具体的设计方案。雷达成像处理平台的软件设计是利用成像处理板卡进行雷达算法的实时成像。首先介绍了基本的距离多普勒雷达成像算法,并结合运动补偿和几何失真校正对该成像算法修正,得到适合于外场恶劣环境的多分辨率多模式的雷达实时成像算法。然后详细分析了基于TMS320C6678的多核算法映射和任务分配,在充分利用板卡实际资源的前提下,使雷达成像在实时性和分辨率上更加优化。在整个软件设计方面,不仅对程序的整体结构进行合理的安排,也对子函数进行优化处理,子函数的优化包括底层函数的优化,存储空间的合理分配以及同步和Cache一致性维护。最后对算法的时间进行统计分析,在外场实验中得到满足各项指标的成像结果。本文从原理,实际设计和调试这几方面详细分析了雷达成像平台的硬件设计,高速接口调试和算法软件设计,完成后的成像处理平台满足实时成像的各项指标。
[Abstract]:Based on the background of radar imaging, this paper analyzes the actual demand and new development of imaging processing system, and proposes and implements a new synthetic aperture radar imaging platform, which can meet the real-time requirements of synthetic aperture radar imaging. In this design, two pieces of DSP-TMS320C6678 FPGA VPX architecture are used as radar imaging processing platform, and the boards of other functional modules are interlinked with other functional modules through VPX standard interface. The imaging processing platform DSP is TI's new high-performance eight-core DSP-TMS320C6678, with a single core frequency of up to 1.25 GHz. It supports fixed-point and floating-point high-speed operations. It has three independent DMA controllers and rich memory resources. In the aspect of hardware design, the design of power supply and clock module is introduced emphatically. Under the premise of satisfying the power consumption and clock requirement of radar imaging platform, the external memory expansion interface and multi-channel high speed serial interface are introduced. Reduce the power consumption and volume of the card. Next, based on the radar imaging processing platform, The high speed interface on the board is analyzed in detail. The high speed interface of the board includes DDR3 Hyperlink PCIeSIO and Gigabit Ethernet. Firstly, the basic principle and interconnection design of the high speed interface are introduced, and then the configuration and debugging of the interface communication software are analyzed in detail. Finally, the design of high-speed interface speed testing and analysis. In addition, This paper also introduces the general interface of the board such as SPI and GPIO, and puts forward and implements its specific design scheme. The software design of radar imaging processing platform is to use the imaging processing board to carry on the real-time imaging of radar algorithm. Basic range Doppler radar imaging algorithm, Combined with motion compensation and geometric distortion correction, a multi-resolution and multi-mode real-time radar imaging algorithm is obtained, which is suitable for the harsh environment of the external field. Then, the mapping and task assignment of the multi-accounting method based on TMS320C6678 are analyzed in detail. On the premise of making full use of the actual resources of the board, the radar imaging is optimized in real-time and resolution. In the whole software design, not only the overall structure of the program is reasonably arranged, but also the subfunctions are optimized. The optimization of the subfunction includes the optimization of the underlying function, the reasonable allocation of storage space, and the maintenance of synchronization and Cache consistency. Finally, the time of the algorithm is analyzed statistically. In this paper, the hardware design, high speed interface debugging and algorithm software design of radar imaging platform are analyzed in detail from the aspects of principle, actual design and debugging. After the completion of the imaging processing platform to meet the real-time imaging indicators.
【学位授予单位】：西安电子科技大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TN957.52

【相似文献】