Please wait a minute...
 
自然资源遥感  2021, Vol. 33 Issue (3): 107-113    DOI: 10.6046/zrzyyg.2020253
  技术方法 本期目录 | 过刊浏览 | 高级检索 |
面向汛旱情监测的遥感影像GPU并行处理算法
赵晓晨1(), 吴皓楠2, 李林宜1, 孟令奎1()
1.武汉大学遥感信息工程学院,武汉 430079
2.珠江水利委员会珠江水利综合技术中心,广州 510611
GPU-based parallel image processing algorithm for flood and drought monitoring
ZHAO Xiaochen1(), WU Haonan2, LI Linyi1, MENG Lingkui1()
1. School of Remote Sensing and Information Engineering, Wuhan University, Wuhan 430079, China
2. Pearl River Comprehensive Technology and Network Information Center, Guangzhou 510611, China
全文: PDF(3919 KB)   HTML  
输出: BibTeX | EndNote (RIS)      
摘要 

针对面向汛旱情监测应用中遥感影像处理耗时过长的问题,包括辐射校正、几何纠正、遥感指数计算等过程,对其业务化工作流程进行了分解分析。结合统一计算架构(compute unified device architecture,CUDA)的存储结构和程序设计模型,将数据处理过程划分为数据读取、直方图统计、栅格分割、波段计算、重采样和数据输出等模块,对波段计算及重采样等模块设计了并行处理方案,并通过实验确定了栅格划分的最佳尺度,基于栅格数组图形处理器(graphics processing unit,GPU)映射方法加速了数据传输效率,最终提出了基于CUDA架构CPU-GPU协同的并行处理算法。实验结果表明,辐射校正及遥感指数计算的波段计算模块可节约58.9%的时间; 几何纠正效果最为显著,最邻近像元重采样和双线性内插重采样模块的最终加速比分别能够达到9倍和7倍以上。

服务
把本文推荐给朋友
加入引用管理器
E-mail Alert
RSS
作者相关文章
赵晓晨
吴皓楠
李林宜
孟令奎
关键词 汛旱情监测几何纠正遥感指数计算GPUCUDA    
Abstract

Aiming at the time-consuming problems in the remote sensing image processing for flood and drought monitoring, the authors analyzed related workflows and algorithms including radiometric correction, geometric correction, and the calculation of remote sensing indices. Based on the storage structure and program design model of compute unified device architecture (CUDA), the remote sensing image processing was divided into several modules, including data reading, histogram statistics, grid partition, band calculation, resampling, and data output. Among them, parallel processing schemes were designed for the modules of band calculation and resampling, and the optimal cell sizes were determined for the module of grid partition. Meanwhile, the data transfer efficiency was increased through the grid data mapping based on a graphics processing unit (GPU). Finally, a parallel processing algorithm based on CPU-GPU collaboration in CUDA was proposed. The experiment results are as follows. The modules of radiometric correction and band calculation of remote sensing indices showed a 58.9% saving in time. Meanwhile, the geometric correction module enjoyed the most significant time-saving effects, and the final speedup ratios of the resampling methods of nearest neighbor and bilinear interpolation reached up to nine and seven times, respectively.

Key wordsflood and drought monitoring    geometric correction    calculation of remote sensing ndex    GPU    CUDA
收稿日期: 2020-08-17      出版日期: 2021-09-24
ZTFLH:  TP751P237  
基金资助:国家重点研发计划课题“河湖岸线洲滩立体监测及河湖功能动态评估关键技术研究”(2018YFC0407804)
通讯作者: 孟令奎
作者简介: 赵晓晨(1995-),女,硕士研究生,主要从事水利遥感方面的研究。Email: zhaoxiaochen@whu.edu.cn
引用本文:   
赵晓晨, 吴皓楠, 李林宜, 孟令奎. 面向汛旱情监测的遥感影像GPU并行处理算法[J]. 自然资源遥感, 2021, 33(3): 107-113.
ZHAO Xiaochen, WU Haonan, LI Linyi, MENG Lingkui. GPU-based parallel image processing algorithm for flood and drought monitoring. Remote Sensing for Natural Resources, 2021, 33(3): 107-113.
链接本文:  
https://www.gtzyyg.com/CN/10.6046/zrzyyg.2020253      或      https://www.gtzyyg.com/CN/Y2021/V33/I3/107
Fig.1  RPC模型几何纠正算法
Fig.2  遥感影像处理算法并行模式
Fig.3  并行化RPC纠正算法
字段 含义
Name 型号 GeForce GT 640M LE
Compute Capability 计算能力/ TFlops 2.1
Warp Size 线程束内线程数量/个 32
Threads Per Block 线程块最大线程数/个 1 024
Threads Per Multiprocessor 流处理器簇最大线程数/个 1 536
Tab.1  GPU硬件参数
项目
型号 GeForce GT 640M LE
计算能力/ TFlops 2.1
时钟频率/ MHz 1 505
总线带宽/ bits 128
流处理器簇 2
流处理器(CUDA核心) 96
流处理器簇最大线程数/个 1 536
线程束尺寸 32
线程块尺寸 1 024
线程块寄存器数量/个 3 268
驱动版本 376.51
CUDA版本 8.0
CUDA运行库版本 6.5
Tab.2  GPU硬件及CUDA详细参数
Fig.4  线程块尺寸实验
Fig.5  栅格划分尺度实验
Fig.6  指数并行算法实验结果
Fig.7  鄱阳湖多光谱指数
波段号 影像读取 直方图统计 波段运算 产品输出 合计
传统串行 CUDA并行 传统串行 CUDA并行 传统串行 CUDA并行 传统串行 CUDA并行 传统串行 CUDA并行
B1 2.76 3.57 32.71 31.7 16.34 7.54 2.90 3.70 54.71 46.51
B2 3.42 3.29 28.76 40.29 22.90 8.30 3.16 2.21 58.24 54.09
B3 8.35 4.57 40.23 37.75 17.95 10.48 3.21 2.39 69.74 55.19
B4 4.57 7.63 22.95 39.32 14.21 6.54 1.98 3.25 43.71 56.74
B5 6.54 9.47 34.69 30.07 26.30 9.36 4.32 5.73 71.85 54.63
B6 2.32 3.12 32.81 31.60 18.68 10.23 6.31 4.16 60.12 49.11
B7 3.21 2.67 39.65 28.40 19.73 6.98 3.29 5.27 65.88 43.32
B8 7.48 4.32 27.03 38.27 20.10 4.38 2.70 4.90 57.31 51.87
B9 3.24 6.32 26.54 33.91 19.43 7.49 2.40 3.57 51.61 51.29
B10 5.31 2.13 35.63 29.50 27.31 7.90 2.98 4.78 71.23 44.31
B11 3.59 4.36 31.26 35.10 15.49 10.61 2.67 2.90 53.01 52.97
均值 4.62 4.68 32.02 34.17 19.86 8.16 3.27 3.90 59.76 50.91
Tab.3  传统串行和CUDA并行辐射处理耗时对比
Fig.8  RPC校正算法耗时对比
重采样方式 步骤 串行/s 并行/s 加速比
最邻近像元 重采样 174.47 18.79 9.29
总计 195.88 61.96 3.16
双线性内插 重采样 506.32 71.36 7.10
总计 529.60 118.75 4.46
Tab.4  RPC并行校正算法重采样环节加速比
[1] 肖汉, 张祖勋. 基于GPGPU的并行影像匹配算法[J]. 测绘学报, 2010, 39(1):46-51.
Xiao H, Zhang Z X. Parallel image matching algorithm based on GPGPU[J]. Acta Geodaetica et Cartographica Sinica, 2010, 39(1):46-51.
[2] 方留杨, 王密, 李德仁. CPU和GPU协同处理的光学卫星遥感影像正射校正方法[J]. 测绘学报, 2013, 42(5):668-675.
Fang L Y, Wang M, Li D R. A CPU-GPU co-processing orthographic rectification approach for optical satellite imagery[J]. Acta Geodaetica et Cartographica Sinica, 2013, 42(5):668-675.
[3] 张过, 潘红播, 江万寿, 等. 基于RPC模型的线阵卫星影像核线排列及其几何关系重建[J]. 国土资源遥感, 2010, 22(4):1-5.doi: 10.6046/gtzyyg.2010.04.01.
doi: 10.6046/gtzyyg.2010.04.01
Zhang G, Pan H B, Jiang W S. Epipolar resampling and epipolar geometry reconstruction of linear array scanner scenes based on RPC model[J]. Remote Sensing for Land and Resources, 2010, 22(4):1-5.doi: 10.6046/gtzyyg.2010.04.01.
doi: 10.6046/gtzyyg.2010.04.01
[4] McFeeters S K. The use of the normalized difference water index (NDWI) in the delineation of open water features[J]. International Journal of Remote Sensing, 1996, 17(7):1425-1432.
doi: 10.1080/01431169608948714
[5] 徐涵秋. 利用改进的归一化差异水体指数(MNDWI)提取水体信息的研究[J]. 遥感学报, 2005, 1(5):589-595.
Xu H Q. A study on information extraction of water body with the modified normalized difference water index(MNDWI)[J]. Journal of Remote Sensing, 2005, 1(5):589-595.
[6] 马秀丹, 崔宾阁, 钟勇, 等. 一种基于GPU和内存映射文件的高分辨率遥感图像快速处理方法[J]. 海洋科学, 2018, 42(1):139-146.
Ma X D, Cui B G, Zhong Y, et al. A fast processing method for high-resolution remote sensing image based on GPU and memory mapping file[J]. Marine Sciences, 2018, 42(1):139-146.
[7] 沈夏炯, 侯柏成, 韩道军, 等. 基于GPU和矩阵分块的增强植被指数计算[J]. 遥感信息, 2018, 33(3):63-69.
Shen X J, Hou B C, Han D J, et al. Calculation of enhanced vegetation index based on GPU and matrix partition[J]. Remote Sensing Information, 2018, 33(3):63-69.
[8] NVIDIA.CUDA 2.0 for windows CUDA 2.0 programming guide[EB/OL]. [2008-06-07]. http://developer.download.nvidia.com/compute/cuda/20/docs/NVIDIA CUDA Programming Guide 2.0.pdf.20.
[9] Rajpoot P S, Kumar A. Impact assessment of meteorological drought on rainfed agriculture using drought index and NDVI modeling:A case study of Tikamgarh district[J]. Applied Geomatics, 2019, 11(1):15-23.
doi: 10.1007/s12518-018-0230-6
[10] Huang F, Chen Y, Li L, et al. Implementation of the parallel mean shift-based image segmentation algorithm on a GPU cluster[J]. International Journal of Digital Earth, 2019, 12(3):328-353.
doi: 10.1080/17538947.2018.1432709
[11] Heo J M, Jo G, Han H J, et al. Accelerated code generator for processing ocean color remote sensing data on GPU[C]// IEEE.IGARSS 2018 IEEE International Geoscience and Remote Sensing Symposium, 2018:9218-9221.
[12] Nakamura T, Murakami S, Sambo L, et al. Efficacy of a GPGPU-acceleration to inundation flow simulation in tonle sap lake in cambodia[J]. Engineering Journal, 2019, 23(1):151-169.
doi: 10.4186/ej
[13] Zhou H F, Zhao J. Parallel programming design and storage optimization of remote sensing image registration based on GPU[J]. Journal of Computer Research and Development, 2012, 49(1):281-286.
[14] Guo X, Wu J, Wu Z, et al. Parallel computation of aerial target reflection of background infrared radiation:Performance comparison of OpenMP,OpenACC,and CUDA implementations[J]. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2016, 9(4):1653-1662.
doi: 10.1109/JSTARS.4609443
[15] Gegout P, Oberlé P, Desjardins C, et al. Ray-tracing of GNSS signal through the atmosphere powered by CUDA,HMPP and GPUs technologies[J]. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2014, 7(5):1592-1602.
doi: 10.1109/JSTARS.4609443
[16] Norman M, Larkin J, Vose A, et al. A case study of CUDA FORTRAN and OpenACC for an atmospheric climate kernel[J]. Journal of Computational Science, 2015, 9(1):1-6.
doi: 10.1016/j.jocs.2015.04.022
[17] Peña A J, Reaño C, Silla F, et al. A complete and efficient CUDA-sharing solution for HPC clusters[J]. Parallel Computing, 2014, 40(10):574-588.
doi: 10.1016/j.parco.2014.09.011
[1] 马御棠, 潘浩, 周仿荣, 黄然, 赵加能, 罗继强, 刘靖, 孙浩轩, 贾伟洁, 张涛. 基于卫星影像的输电杆塔坐标校准方法[J]. 自然资源遥感, 2022, 34(2): 63-71.
[2] 杨亮, 贾益, 江万寿, 张过. 基于观测角信息的HJ-1A/B卫星光学影像几何精纠正[J]. 国土资源遥感, 2018, 30(2): 60-66.
[3] 李朋龙, 丁忆, 胡艳, 罗鼎, 段松江, 舒文强. 一种基于GPU并行计算的无人机影像快速镶嵌方法[J]. 国土资源遥感, 2017, 29(4): 57-63.
[4] 江威, 何国金, 龙腾飞, 尹然宇, 宋小璐, 袁益琴, 凌赛广. 基于北斗和GPS的高分二号全色影像正射精度验证与分析[J]. 国土资源遥感, 2017, 29(3): 211-216.
[5] 贾益, 王盛, 江万寿. 中低分辨率卫星影像Bowtie效应的快速去除[J]. 国土资源遥感, 2016, 28(4): 83-87.
[6] 马利春, 杨仁忠, 石璐. Forman相位校正算法的改进与实现[J]. 国土资源遥感, 2013, 25(3): 97-101.
[7] 何敬, 李永树, 鲁恒, 张帅毅. 无人机影像地图制作实验研究[J]. 国土资源遥感, 2011, 23(4): 74-77.
[8] 万紫, 徐茂松, 夏忠胜, 张风丽, 宫华泽. 高分辨率雷达图像双视向几何纠正方法研究[J]. 国土资源遥感, 2010, 22(2): 12-16.
[9] 宋薇. CBERS-02星图像几何纠正方法试验研究[J]. 国土资源遥感, 2009, 21(1): 51-54.
[10] 李朝奎, 周国清. 基于DLT算法的微型无人机(MUAV)视频影像的几何纠正方法[J]. 国土资源遥感, 2006, 18(4): 23-28.
[11] 何维, 李秉柏, 张娅香, 金之庆. 一种新的控制点采集方法[J]. 国土资源遥感, 2001, 13(4): 59-63.
[12] 邸凯昌. 微机遥感图像几何纠正快速算法[J]. 国土资源遥感, 1996, 8(3): 45-48.
Viewed
Full text


Abstract

Cited

  Shared   
  Discussed   
京ICP备05055290号-2
版权所有 © 2015 《自然资源遥感》编辑部
地址:北京学院路31号中国国土资源航空物探遥感中心 邮编:100083
电话:010-62060291/62060292 E-mail:zrzyyg@163.com
本系统由北京玛格泰克科技发展有限公司设计开发