摘要
针对实际中所观测到的探地雷达信号时常会被噪声污染的问题,根据小波分析和分形理论在多尺度分析和自相似本质上的一致性,从研究分数布朗运动小波系数的统计特性,分析信号和噪声对小波系数统计特性的影响,在探地雷达信号处理中引入小波分析和随机分形理论,提出了在小波域中,使用平滑因子恢复加性白噪背景下探地雷达信号。仿真实验表明,该方法能有效地恢复白噪背景下探地雷达信号,显著地提高了信噪比(由0 db提高到13.9 db)。
Considering that the observed signals in ground penetrate radar data are sometimes polluted by noise,we introduce,in this paper,the application of smoothing factor in wavelet domain to estimate available signal from GPR data interfered by white noise,according to wavelet analysis and fractal theory both havng identical properties in multi-scale analysis and self similarity.First,the signal with white noise is transformed via the wavelet analysis filter into multi-scale subsystems in time scale domain.Some non-stationary properties of fractal signals are attenuated in each scale by wavelet multi-resolution decomposition,then smoothing factor can be applied to estimate the multi-scale signals.Finally the estimated multi-scale signals are synthesized to obtain the estimated signal.With this multi-scale analysis/ synthesis,this method can extract available signal of 1/ f family of fractal from synthesized signal with white noise.A simulation examples is given for testing the performance of the proposed algorithm. The signal-to-noise ratio is evidently increased form 0 db to 13.9db.
出处
《物探化探计算技术》
CAS
CSCD
2006年第4期352-356,共5页
Computing Techniques For Geophysical and Geochemical Exploration
基金
湖北省自然科学基金项目(2005ABA275)
关键词
分形
分数布朗运动
1/f类信号
小波分析
探地雷达
fractal
fractional Brownian motion
1/ f family signal
wavelet transform
ground penetrate radar
