系统仿真学报 ›› 2017, Vol. 29 ›› Issue (1): 120-128.doi: 10.16182/j.issn1004731x.joss.201701017

• 仿真系统与技术 • 上一篇    下一篇

基于Hadoop的探地雷达数据并行处理方法研究

梁胤程, 袁媛, 杨峰   

  1. 中国矿业大学(北京) 机电与信息工程学院,北京 100083
  • 收稿日期:2015-04-27 修回日期:2015-07-10 出版日期:2017-01-08 发布日期:2020-06-01
  • 作者简介:梁胤程(1983-),男,山东,博士生,研究方向为大数据处理;袁媛(1990-),女,河北,硕士,研究方向为并行计算;杨峰(1968-),男,河南,教授,研究方向为探地雷达仪器开发。
  • 基金资助:
    国家重大科学仪器设备开发专项(2012YQ 030126),核三废专项科研课题(环FZ1402-3)

Research of Parallel Process Method of Ground Penetrating Radar (GPR) Data Based on Hadoop

Liang Yincheng, Yuan Yuan, Yang Feng   

  1. School of Mechanical Electronic&Information Engineering, China University of Mining&Technology, Beijing 100083, China
  • Received:2015-04-27 Revised:2015-07-10 Online:2017-01-08 Published:2020-06-01

摘要: 大数据及云计算等技术带来的科研方式的转变影响着探地雷达数据处理领域的研究工作,利用云计算平台对大规模探地雷达数据进行了数据解析的研究工作,通过流式处理框架Storm平台对产生的探地雷达数据进行了数据预处理,解决了探地雷达数据格式与Hadoop平台数据处理输入格式不匹配的问题。并在Hadoop平台下实现了并行化的卷积、反卷积与增益相结合的数据滤波处理,对滤波效果和运行性能指标进行了分析。仿真实验结果表明,该方法能准确有效地对探地雷达数据进行滤波解析,在运算速度、并行化加速度性能上都较传统方法有明显的改善。

关键词: Hadoop, 探地雷达数据, Storm, 卷积, 增益, 并行化

Abstract: The change of research way brought by big data and cloud computing has an impact on the research of GPR data processing. Filtering and analysis research for large scale GPR data was conducted by cloud computing platforms. Data preprocessing to GPR data was realized through flow computing framework Storm so as to solve the mismatch of GPR data format and input data format of Hadoop and realize parallel convolution and gain process based on Hadoop. The filtering result and performance index was analyzed. Experimental results show that the proposed method is accurate and effective on GPR data process, and it has significantly improved operating speed and parallel speed-up in comparison with previous method.

Key words: Hadoop, GPR data, Storm, convolution, gain process, parallel

中图分类号: