地球物理学进展 ›› 2019, Vol. 34 ›› Issue (5): 2097-2105.doi: 10.6038/pg2019CC0351

• 应用地球物理学Ⅱ(海洋、工程、环境、仪器等) • 上一篇    下一篇

小波与曲波变换探地雷达数据去噪对比分析

杨俊,李静和,孟淑君,廖小倩,李文杰   

  1. 桂林理工大学地球科学学院,广西桂林 541004
  • 收稿日期:2018-11-13 修回日期:2019-05-27 出版日期:2019-10-28 发布日期:2019-10-28
  • 作者简介:杨俊, 男, 1992生, 硕士研究生, 主要从事数值模拟与反演研究. (E-mail: yjun_gujie@163.com)
  • 基金资助:
    广西自然科学基金项目(2016GXNSFBA380195);广西自然科学基金项目(2018JJA150001)

Comparative analysis of ground penetrating radar data denoising with wavelet transform and curvelet transform

YANG Jun,LI Jing-he,MENG Shu-jun,LIAO Xiao-qian,LI Wen-jie   

  1. College of Earth Sciences, Guilin University of Technology, Guangxi Guilin 541004, China
  • Received:2018-11-13 Revised:2019-05-27 Online:2019-10-28 Published:2019-10-28

摘要:

探地雷达勘探工程目标及观测工程环境越加复杂给其精确的数据处理带来极大挑战,高效的探地雷达数据去噪算法是当前关注的重要研究领域.基于阈值去噪思想,小波变换和曲波变换去噪算法在探地雷达数据去噪应用中受到限制,有必要开展上述两种去噪算法适用性和实用性系统评价及改进.基于传统高阶相关统计阈值和块状复数域阈值函数,本文开展了小波变换及曲波变换去噪算法在合成含噪数据去噪效果对比分析;提出窗口高阶相关统计阈值小波变换去噪算法、探讨了块状复数域阈值函数取值变化对曲波变换去噪效果影响规律.通过对实测数据去噪分析,验证了窗口高阶相关统计阈值小波变换和估计块状复数域阈值函数曲波变换去噪算法的可行性及有效性.

关键词: 探地雷达, 小波变换, 曲波变换, 阈值函数, 对比分析

Abstract:

The complexity of the exploration targets and the observation environment has brought great challenges to the precise GPR data processing. Among them, the effective ground penetrating radar data denoising algorithm is an important research area. Based on the idea of threshold denoising, the wavelet transform and curvelet transform denoising algorithm are limited in the application of GPR data denoising. It is necessary to carry out the applicability and practical evaluation of the two denoising algorithms mentioned above. In this paper, based on the traditional high order correlation statistical threshold and the block complex domain threshold function, the wavelet transform and the curvelet transform denoising algorithm are compared and analyzed. A ‘window’ high order correlation statistical threshold wavelet transform denoising algorithm is proposed and the denoising effect of the block complex domain threshold function is discussed, and also discussing the denoising effect of the curvelet transform. Through denoising analysis of the measured data, the feasibility and effectiveness of the high order correlation statistical threshold wavelet transform and the estimation of the curvelet transform denoising algorithm of the block complex domain threshold function are verified.

Key words: Ground penetrating radar, Wavelet transform, Curvelet transform, Threshold function, Comparative analysis

中图分类号: