| 一种分块相位梯度自聚焦算法的并行实现 |
| |
| 引用本文： | 叶春茂,宁夏,杨健,樊康. 一种分块相位梯度自聚焦算法的并行实现[J]. 清华大学学报(自然科学版), 2012, 0(5): 612-615,622 |
| |
| 作者姓名： | 叶春茂 宁夏 杨健 樊康 |
| |
| 作者单位： | 清华大学电子工程系;中国人民解放军66222部队 |
| |
| 基金项目： | 部委项目航空科学基金(20102058014) |
| |
| 摘 要： | 为有效实现合成孔径雷达(SAR)平台运动误差的补偿,提出了一套适合实际应用的分块相位梯度自聚焦(PGA)并行实现算法。首先,针对大场景观测中运动误差的空变性,提出了分块的PGA方法;然后,针对PGA补偿中所涉及较大运算量,采用通用图形计算显卡(GPGPU)进行了算法的并行化实现。采用NVIDIA公司的Tesla C2050对所述方法进行了实现,取得了良好的运动补偿效果以及优异的处理速度提升。该研究结果为SAR成像处理器的设计提供了一种有效途径。
|
| 关 键 词： | 合成孔径雷达 相位梯度自聚焦 空变性 并行计算 |
Parallel implementation of a block-wise phase gradient autofocusing method |
| |
| YE Chunmao,NING Xia,YANG Jian,FAN Kang. Parallel implementation of a block-wise phase gradient autofocusing method[J]. Journal of Tsinghua University(Science and Technology), 2012, 0(5): 612-615,622 |
| |
| Authors: | YE Chunmao NING Xia YANG Jian FAN Kang |
| |
| Affiliation: | 1.Department of Electronic Engineering,Tsinghua University, Beijing 100084,China; 2.Department 66111,People’s Libration Army, Beijing 100042,China) |
| |
| Abstract: | A parallelized implementation of a block-wise phase gradient autofocusing(PGA) method is given in this paper that effectively compensates for the motion error induced by the platform for synthetic aperture radar(SAR).A block-wise PGA method is used to overcome the space-variant characteristics of the motion error when observing a wide scene.Then,a parallelized implementation of this PGA method is realized on a general purpose graphic processing unit(GPGPU) to handle the large computational load.Successful implementation on a NVIDIA Tesla C2050 provides a well-focused image with a large speedup ratio.The results show that the method is effective for SAR processors. |
| |
| Keywords: | synthetic aperture radar(SAR) phase gradient autofocusing(PGA) space-variant characteristic parallel computation |
| 本文献已被 CNKI 等数据库收录！ |
|
