SAR实时成像系统的快速定点仿真验证技术  

Fast Fixed-Point Simulation Technology for SAR Real-Time Imaging System

作　　者：李炳沂 谢宜壮[1] 杜青[1] 陈禾[1] 杨晨[1] 

机构地区：[1]北京理工大学信息与电子学院,北京100081

出　　处：《北京理工大学学报》2017年第1期67-70,76,共5页Transactions of Beijing Institute of Technology

摘　　要：SAR实时成像处理系统的定点化优化能够减小系统存储规模、提高系统运算速度,是降低系统功耗、减小系统规模、提高实时性的有效方法.本文以CS成像算法为例,分析并统计了算法的运算量,针对其核心运算——FFT进行了定点化优化,选择基-22FFT算法,设计了合理的定标策略.采用SystemC语言对整个SAR成像系统进行了系统级快速仿真和验证,真实模拟了FPGA系统的实现结果.对16 384×16 384大小的点阵数据和面目标数据采用24bits的字长进行了成像仿真,仿真结果满足系统指标要求.Fixed-point processing optimization can reduce the storage,improve the computing speed,is an effective way to reduce power consumption and storage,improve the real-time performance of synthetic aperture radar（SAR）real-time imaging system.Taking the well-known chirp scaling（CS）algorithm as example,the count of computations was analyzed in this paper.Radix-22 algorithm of fast Fourier transform（FFT）as the main part of computations was adopted for fixed-point processing optimization,realized through a reasonable scale strategy with SystemC.Thus rapid system-level simulation of CS algorithm was implemented to provide a standard basis for FPGA implementation.Finally,test images from the raw lattice and area target data（16 384×16 384）were analyzed using fixed-point CS algorithm with 24 bits input word lengths.The simulation results meet the requirements of system indicators.

关 键 词：合成孔径雷达(SAR) 实时成像系统 定点化 仿真验证 SYSTEMC 

分 类 号：TN47[电子电信—微电子学与固体电子学]