太赫兹科学与电子信息学报, 2019, 17 (1): 146, 网络出版: 2019-04-07
基于CPU+GPU混合架构的实时成像系统设计与实现
Design and implementation of a real-time imaging system based on CPU+GPU hybrid architecture
图形处理器 混合架构 合成孔径雷达 逆合成孔径雷达 实时成像系统 电子战 Graphics Processing Unit hybrid architecture Synthetic Aperture Radar Inverse Synthetic Aperture Radar real-time imaging system electronic warfare
摘要
雷达成像处理需要更大宽带以实现更高的距离分辨力,同时还需要更多的脉冲积累获得更高的方位像分辨力，因此雷达成像处理过程计算量巨大。如何实现未来超带宽雷达的实时成像处理是一项艰巨挑战。图形处理器(GPU)以卓越的浮点性能和访存带宽，成为并行加速应用平台的有力候选者。设计了一种基于CPU+GPU平台并面向合成孔径雷达/逆合成孔径雷达(SAR/ISAR)的实时成像系统方案，并将该方案实体化。实验表明，该成像系统能够实现实时SAR/ISAR成像，同时该实时成像系统也可用于电子对抗领域，在干扰方法和效果研究中起到重要作用。
Abstract
Radar image processing is a radar signals process with huge computation quantity, because it requires wider bandwidth and more pulse accumulation, and it becomes a big challenge to achieve real- time image process for ultra wide band in the future. Graphics Processing Unit(GPU) is the candidate for its excellent computing capability and memory access performance. A real-time Synthetic Aperture Radar/Inverse Synthetic Aperture Radar(SAR/ISAR) imaging system based on CPU+GPU hybrid architecture is designed and implemented. The results show that the real-time SAR/ISAR system can achieve wide band SAR/ISAR image process, and it can be applied in the electronic warfare field and play a key part in the jammer study and experiment.
张彦彬, 丁晟, 高雁, 李松键, 朱金中, 孙友礼, 张陨石. 基于CPU+GPU混合架构的实时成像系统设计与实现[J]. 太赫兹科学与电子信息学报, 2019, 17(1): 146. ZHANG Yanbin, DING Sheng, GAO Yan, LI Songjian, ZHU Jinzhong, SUN Youli, ZHANG Yunshi. Design and implementation of a real-time imaging system based on CPU+GPU hybrid architecture[J]. Journal of terahertz science and electronic information technology, 2019, 17(1): 146.
PDF全文
