首页 > 论文 > 光电工程 > 43卷 > 3期(pp:66-72)

相位差算法在多 GPU平台上的并行化实现

Parallel Implementation of Phase Diversity Algorithm on Multi-GPU

  • 摘要
  • 论文信息
  • 参考文献
  • 被引情况
  • PDF全文
分享:

摘要

相位差算法 (PD)不仅应用于波前传感器还可以作为一种图像复原技术,在波前复原与高分辨力图像事后图像复原有很大的潜力;但是,由于相位差算法的计算量大导致了在 PC平台上很难实现实时处理。以 CPU与 GPU之间较少的数据交互为原则,本文首先对 PD计算过程进行了分析,然后在多 GPU平台上,对 PD算法进行任务划分及优化。实验结果表明,在相差 PV=2λ条件下对于 256 pixels×256 pixels大小图像,经过 50次迭代单 GPU和双 GPU分别耗时 53 ms和 45 ms。

Abstract

Phase diversity (PD) can not only be used as wave-front sensor but also as image restoration technique. However, its computations have been perceived as being too burdensome and it is not to satisfy the real-time requirement. On the principle of minimum data communication between CPU and GPU, we analyze the process of PD algorithm and the experiment result shows that the part of optimization is appropriate to be realized on GPU. Firstly, the process of PD is analyzed and then task partition is implemented on dual GPUs. Under the experiment condition of wavefront aberration of PV=2λ, after 50 iterations, the execution time of single GPU and dual GPUs are 53 ms and 45 ms respectively for image size of 256 pixels×256 pixels.

Newport宣传-MKS新实验室计划
补充资料

中图分类号:TP391

DOI:10.3969/j.issn.1003-501x.2016.03.011

所属栏目:图像与信号处理

基金项目:国家自然科学基金( 11178004);中国科学院实验室创新基金( YJ14K018)

收稿日期:2015-02-04

修改稿日期:2015-06-12

网络出版日期:--

作者单位    点击查看

张全:中国科学院自适应光学重点实验室,成都 610209电子科技大学光电信息学院,成都 610054中国科学院光电技术研究所,成都 610209中国科学院大学,北京 100049
鲍华:中国科学院自适应光学重点实验室,成都 610209中国科学院光电技术研究所,成都 610209
饶长辉:中国科学院自适应光学重点实验室,成都 610209中国科学院光电技术研究所,成都 610209
彭真明:电子科技大学光电信息学院,成都 610054

联系人作者:张全(quanzhang100@126.com)

备注:张全 (1985-),男(汉族),甘肃武威人。博士研究生,主要研究工作是 GPU高性能计算,图像处理。

【1】Georges Ⅲ J A,Dorrance P,Gleichman K,et al. High-speed closed-loop dual deformable-mirror phase-diversity testbed [C]// Advanced Wavefront Control:Methods,Devices,and Applications V,San Diego,California,USA,August 26,2007: 671105.

【2】Warmuth M W,Parker S W,Wilson A J,et al. Operation of phase-diverse adaptive-optics with extended scenes [C]// Optical Engineering Applications. International Society for Optics and Photonics,San Diego,California,USA,August 10,2008: 709307.

【3】赵金宇,陈占芳,王斌,等 . 相位差异法目标函数的并行化改造 [J].光学精密工程, 2012,20(2):431-438. ZHAO Jinyu,CHEN Zhanfang,WANG Bin,et al. Parallelity improvement of object function for phase diversity [J]. Optics and Precision Engineering,2012,20(2):431-438.

【4】张楠. 基于相位差异的地基望远镜图像恢复算法与 GPU高速实现 [D].长春:中国科学院长春光学精密机械与物理研究所,2012:59-77. ZHANG Nan. Image restoration algorithm based on phase diversity for ground-based telescope and high speed implementation on GPU [D]. Changchun:Changchun Institute of Optics,Fine Mechanics and Physics,Chinese Academy of Science,2012: 59-77.

【5】杨磊. Phase diversity波前重构的研究及在高分辨图像复原中的应用 [D].昆明:中国科学院云南天文台, 2007:1-10. YAN Lei. Researches of the phase diversity wave-front reconstruction techniques and its application in the high resolution image restoration [D]. Kunming:Yunnan Observatory of Chinese Academy of Sciences,2007:1-10.

【6】李斐. 相位差法波前探测技术及其在图像恢复中的应用研究 [D].长沙:国防科学技术大学, 2011:30-43. LI Fei. Phase diversity wavefront sensing and its application in image restoration [D]. Changsha:National University of Defense Technology,2011:30-43.

【7】汪宗洋,王建立,王斌,等 . 基于相位差异的图像复原方法 [J].光电工程, 2010,37(12):25-29. WANG Zongyang,WANG Jianli,WANG Bin,et al. Image restoration based on phase diversity [J]. Opto-Electronic Engineering,2010,37(12):25-29.

【8】陈宝林. 最优化理论与方法 [M].北京:清华大学出版社, 2005:281-315,394-412. CHEN Baolin. Theory and Methods of Optimization [M]. Beijing:Tsinghua University Press,2005:281-315,394-412.

【9】Zielinski T P. Robust Image-based Wavefront Sensing [M]. University of Rochester,2011:14-16.

【10】Volkov V,Kazian B. Fitting FFT onto the G80 architecture [EB/OL]// [2014-10]. http://www.cs.berkeley.edu/~kubitron/courses /cs258-S08/ projects/reports/ project6_report.pdf.

引用该论文

ZHANG Quan,BAO Hua,RAO Changhui,PENG Zhenming. Parallel Implementation of Phase Diversity Algorithm on Multi-GPU[J]. Opto-Electronic Engineering, 2016, 43(3): 66-72

张全,鲍华,饶长辉,彭真明. 相位差算法在多 GPU平台上的并行化实现[J]. 光电工程, 2016, 43(3): 66-72

您的浏览器不支持PDF插件,请使用最新的(Chrome/Fire Fox等)浏览器.或者您还可以点击此处下载该论文PDF