一种子孔径ωK聚束SAR成像算法的GPU实现  

GPU Realization of A SAR Imaging Algorithm for Sub-apertureωK Spotlight

作　　者：司军[1] 陈家瑞[1] SI Jun;CHEN Jia-rui(The 723 Institute of CSIC,Yangzhou 225101,China)

机构地区：[1]中国船舶重工集团公司第七二三研究所,江苏扬州225101

出　　处：《舰船电子对抗》2021年第1期59-62,共4页Shipboard Electronic Countermeasure

摘　　要：合成孔径雷达(SAR)的数据运算量不断增加,图形处理器(GPU)为其处理提供了新的运算平台。但是GPU显存小,不足以容纳大场景SAR数据。通过研究聚束SAR成像模式特点,提出了一种适合GPU加速的子孔径成像方案,降低了该算法对GPU显存的要求。在Tesla C2075上的实验结果表明,该方案能够取得良好的成像效果,与CPU上的处理效率相比,有数10倍的速度提升。The amount of synthetic aperture radar(SAR)data calculation is increasing constantly,and the graphic processing unit(GPU)offers a new operation platform for SAR imaging processing.However,GPU memory is not large enough to accommodate the SAR data of large scene.By studying the characteristics of spotlight SAR imaging mode,this paper puts forward a sub-aperture imaging solution which is suitable for GPU acceleration,and the scheme reduces the requirements of the algorithm to memory.The experiment result performed on Tesla C2075 GPU platform shows that this scheme can obtain a good imaging result and accelerate the SAR data processing by tens of times compared with the processing efficiency on CPU.

关 键 词：子孔径 波数域成像算法 图形处理器 聚束模式 

分 类 号：TN957.52[电子电信—信号与信息处理]