吉林大学学报(地球科学版) ›› 2022, Vol. 52 ›› Issue (3): 775-784.doi: 10.13278/j.cnki.jjuese.20210276

• 第十五届中国国际地球电磁学术研讨会专栏 • 上一篇    下一篇

基于卷积神经网络的地下水磁共振数据随机噪声压制方法

李邦1,蒋川东1,2,3,王远1,2,3,田宝凤1,2,段清明1,2,3,尚新磊1,2,3   

  1. 1.吉林大学仪器科学与电气工程学院,长春130026

    2.地球信息探测仪器教育部重点实验室(吉林大学),长春130026

    3.国家地球物理探测仪器工程技术研究中心(吉林大学),长春130026

  • 出版日期:2022-05-26 发布日期:2024-01-03
  • 基金资助:

    吉林省自然科学基金项目(20190201111JC);吉林省教育厅科学研究项目(JJKH20211052KJ,JJKH2021087KJ)



Random Noise Suppression for Groundwater Magnetic Resonance Sounding Data Based on Convolutional Neural Network

Li Bang1, Jiang Chuandong 1,2,3,Wang Yuan1,2,3, Tian Baofeng1,2, Duan Qingming1,2,3, Shang Xinlei1,2,3   

  1. 1. College of Instrumentation & Electrical Engineering, Jilin University, Changchun 130026, China

    2. Key Laboratory of Geophysical Exloration Equipment (Jilin Universty), Minstry of Education of China, Changchun

     130026, China

    3. National Geophysical Exploration Equiipment Enginerring Research Center (Jilin University), Changchua 130026, China

  • Online:2022-05-26 Published:2024-01-03
  • Supported by:
    Supported by the National Natural Science Foundation of Jilin Province (20190201111JC) and the Scientific Research Project of Jilin Provincial Education Department (JJKH20211052KJ,JJKH2021087KJ)

摘要: 磁共振探测(magnetic resonance sounding, MRS)是一种直接探测地下水的地球物理方法,具有定量、准确和高效等优点,广泛应用于水资源调查等领域。MRS信号的质量对于磁共振数据的解释具有重要作用。本文针对强环境干扰情况下MRS信号的随机噪声压制问题展开研究。基于卷积神经网络(convolutional neural networks,CNN)框架,采用监督学习的训练方式,得到含噪信号的时频谱与原始无噪声信号的时频谱间的非线性映射关系,进而实现磁共振信号噪声的压制。仿真结果表明,CNN方法对含噪MRS信号的信噪比提升可达15 dB以上。并对比分析了CNN方法与时频峰值滤波(time-frequency peak filtering,TFPF)方法的噪声压制效果,证明了该方法的有效性和优越性。最后,使用该方法对野外实测数据的噪声压制,进一步证明了本方法的有效性和实用性。

关键词: 地下水, 磁共振探测, 卷积神经网络, 噪声压制

Abstract:

Magnetic resonance sounding (MRS) is a geophysical method for directly detecting groundwater. It is widely used in water resources investigation and other fields, as it is quantitative, accurate, and efficient. The quality of MRS signal strongly affects the interpretation of magnetic resonance data. In this paper, the random noise suppression of MRS signal under strong environmental interference is studied. Based on the framework of convolutional neural networks (CNN), the nonlinear mapping relationship between the time spectrum of noise signal and the time spectrum of original noiseless signal is obtained by using the training method of supervised learning, so as to suppress the noise of magnetic resonance signal. The simulation results show that the signal-to-noise ratio of noisy MRS signal can be improved by more than 15 dB. The noise suppression effect of CNN is analyzed and compared with that of the time-frequency peak filtering (TFPF) method. The effectiveness and superiority of this method is proved by the noise suppression of the measured field data. 

Key words: groundwater, magnetic resonance sounding, convolutional neural networks, noise suppression

中图分类号: 

  • P631.2
[1] 刘 霞, 孙英杰. 基于融合残差注意力机制的卷积神经网络地震信号去噪[J]. 吉林大学学报(地球科学版), 2023, 53(2): 609-.
[2] 张岩, 刘小秋, 李杰, 董宏丽, . 基于时频联合深度学习的地震数据重建[J]. 吉林大学学报(地球科学版), 2023, 53(1): 283-296.
[3] 董旭日, 冯晅, 刘财, 田有, 李静, 王天琪, 王鑫, 衣文索. 基于DAS信号和CNN分类算法的人员运动轨迹监测方法[J]. 吉林大学学报(地球科学版), 2022, 52(3): 1004-1015.
[4] 李忠潭, 薛林福, 冉祥金, 李永胜, 董国强, 李玉博, 戴均豪. 基于卷积神经网络的智能找矿预测方法:以甘肃龙首山地区铜矿为例[J]. 吉林大学学报(地球科学版), 2022, 52(2): 418-433.
[5] 熊越晗, 刘东燕, 刘东升, 王艳磊, 唐小山. 基于岩样细观图像深度学习的岩性自动分类方法[J]. 吉林大学学报(地球科学版), 2021, 51(5): 1597-1604.
Viewed
Full text


Abstract

Cited

  Shared   
  Discussed   
No Suggested Reading articles found!