基于二维小波变换的随机噪声压制方法在GPR数据中的应用  

Application of random noise suppression method based on two-dimensional wavelet transform in GPR data

作　　者：鲁光银[1] 罗帅 朱自强[1] 赵云威[1] 席飞雁 LU Guangyin;LUO Shuai;ZHU Ziqiang;ZHAO Yunwei;XI Feiyan(Institute of Applied Geophysics,School of Geosciences and Info-physic,Central South University,Changsha 410083,China)

机构地区：[1]中南大学地球科学与信息物理学院,长沙410083

出　　处：《物探化探计算技术》2019年第2期234-240,共7页Computing Techniques For Geophysical and Geochemical Exploration

基　　金：国家自然科学基金(41174061;41374120)

摘　　要：在探地雷达探测工作中,为了尽可能多的获取回波信息,通常采用宽频带记录,这就不可避免地将各种干扰波也记录下来,其中随机噪声由于其频带较宽,分布于整个数据剖面,常规滤波方法对随机噪声的压制效果往往不佳。由于小波变换具有较强的分频和局部分析能力,根据需要选择合适的小波基函数和去噪方法,可较好地压制随机噪声,提高信噪比。基于二维小波变换理论,提出了采用自适应分层阈值法对探地雷达数据进行压制随机噪声的方法。通过不同模型、不同信噪比下正演模拟数据的验证以及对实测数据的处理,并与中值滤波法和全局阈值法的去噪效果进行了对比分析,结果表明自适应分层阈值法去噪效果更好,实用性更强。In the ground penetrating radar detection work,in order to obtain as much as possible echo information,broadband records usually is used.This records various disturbances inevitably.Among them,the random noise,due to its wide band is distributed in the entire data profile.The suppression effect of conventional filtering method is often poor.Wavelet transform has strong frequency division and local analysis ability.According to the need to choose the appropriate wavelet basis function and denoising method,it can be better to suppress random noise,improve signal to noise ratio.Based on the theory of two-dimensional wavelet transform,adaptive layered threshold method is adopted to suppress the random noise of ground penetrating radar data.The verification of the different model,forward data of different signal and noise ratio,the processing of the actual data and the comparison with the median filter method and the global threshold method show that the suppression effect for random noise of adaptive layered threshold method is better and more practical.

关 键 词：探地雷达 二维小波变换 随机噪声 自适应分层阈值法 信噪比 

分 类 号：P631.4[天文地球—地质矿产勘探]