• 中国计算机学会会刊
  • 中国科技核心期刊
  • 中文核心期刊

J4 ›› 2015, Vol. 37 ›› Issue (07): 1258-1264.

• 论文 • 上一篇    下一篇

CPU-GPU协同计算的遥感仿真图像MTF退化并行算法

赵瑞斌1,赵生慧1,胡新礼2   

  1. (1.滁州学院计算机与信息工程学院,安徽 滁州239000;2.中国科学院遥感与数字地球研究所,北京100101)
  • 收稿日期:2014-10-15 修回日期:2014-12-20 出版日期:2015-07-25 发布日期:2015-07-25
  • 基金资助:

    安徽省高校自然科学基金重点项目(KJ2014A184);安徽省自然科学基金资助项目 (1408085MF126);安徽省高校自然科学基金资助项目(KJ2013B183)

A CPU-GPU collaboration based computing parallel algorithm
for MTF degradation of remote sensing simulation images  

ZHAO Ruibin1,ZHAO Shenghui1,HU Xinli2   

  1. (1.School of Computer and Information Engineering,Chuzhou University,Chuzhou 239000;
    2.Institute of Remote Sensing and Digital Earth,Chinese Academy of Sciences,Beijing 100101,China)
  • Received:2014-10-15 Revised:2014-12-20 Online:2015-07-25 Published:2015-07-25

摘要:

在遥感图像仿真中,为了定量模拟并分析平台抖动、探测器电子特性、大气衰减等因素对遥感成像质量的影响,需要有效计算遥感系统的调制传递函数MTF,并将其快速作用到仿真图像上。然而,由于遥感仿真图像的大数据量特性以及MTF退化包含多个计算密集型算法,使得计算效率成为一个瓶颈问题。为此,根据已有研究提出的MTF计算模型,分析了遥感仿真图像MTF退化的一般流程及主要环节的算法复杂度。在此基础上,提出了一种CPUGPU协同计算的遥感仿真图像MTF退化并行算法。实验结果表明,该并行算法有效地发挥了GPU并行计算能力,并明显提高了MTF退化处理效率。

关键词: 遥感仿真图像, MTF退化, 并行计算, GPU, CUDA

Abstract:

In order to quantitatively simulate and analyze the impact on the quality of the remote sensing system from factors such as platform jitter,electronic properties,and atmospheric attenuation,it is necessary to compute the modulate transfer function (MTF) of the remote sensing system and operate it in simulation images.However,because of the characteristics of big data in remote sensing image simulation and a number of intensive algorithms involved in the computing of MTF degradation,calculating efficiency becomes the bottleneck problem. Thus, according to the existing MTF calculation model, we analyze the general process of the MTF degradation of remote sensing simulated images and the complexity of the main steps in the algorithm. Based on this,we propose a CPUGPU collaboration based computing parallel algorithm for the MTF degradation of remote sensing simulation images. Experimental results show that the algorithm can make full use of the parallel computing capacity of GPUs and improve the computation efficiency of the MTF degradation.

Key words: remote sensing simulation images;MTF degradation;parallel algorithm;GPU;CUDA