基于小波变换的雷达低信噪比信号处理技术及应用研究  

GPR signal processing approach under low signal to noise ratio based on wavelet transforms and its application

作　　者：柳刚[1] 李术才[1] 薛翊国[1] 单治钢[2] 张霄[1] 

机构地区：[1]山东大学岩土与结构工程中心,济南250061 [2]华东勘测设计研究院,杭州310014

出　　处：《工程勘察》2009年第9期85-90,共6页Geotechnical Investigation & Surveying

基　　金：国家自然科学基金(50539080);国家863计划资助项目(2007AA11Z131);港澳青年学者合作研究基金(50729904)

摘　　要：在分析小波和小波变换特点及小波变换的降噪原理的基础上,基于小波时频局部化优异特性,利用小波分解和软阈值处理方法对实际探地雷达信号进行了降噪处理。通过小波变换信号处理的仿真计算,比较了单道雷达信号去噪处理前后的效果。同时,将小波变换的雷达信号处理剖面和原始剖面进行比较,结果显示经过小波变换的雷达剖面分辨率明显改善。理论计算和实际剖面对比,均表明小波变换较传统方法在处理低信噪比的瞬变信号时具有优势。最后将该法用于锦屏二级水电站交通辅助隧洞的探地雷达超前预报中,发现处理剖面和地质编录的情况基本能够吻合,即小波变换处理剖面有助于在隧道超前预报的正确判释。The basic principle of wavelet transform is introduced. The wavelet transform has an excellent performance in signal processing, especially in unbalanced signal because it could get good features in time and frequency domain. Based on an engineering project, the best wavelet function to help the GPR signal processing according to the unique feature of GPR signal is introduced. Compared the signal from wavelet transform with the results from the professional software, it is obviously that the quality of GPR profile is optimism and the precision of the prediction has been improved. While the processing, it is observed that the frequency domain range could be computed according the lowest level profile. The distribution of the spectrum of the profile reflects the level of noise. It is concluded that the discrete wavelet transform could be well used in GPR signal under low signal to noise ratio （SNR） processing.

关 键 词：小波变换 探地雷达 信号处理 超前预报 

分 类 号：O319.56[理学—一般力学与力学基础]