摘要
地震勘探中的随机噪声一直是影响地震信号信噪比的主要因素之一。本文针对一维经典中值滤波器对地震信号去噪能力过强而导致有用信息被破坏的缺点,基于噪声特性设定一个窗口长度控制阈值,再根据地震数据与阈值的关系选取合适的时变窗口进行中值滤波,以取得有效信号与不同强度噪声之间的最佳平衡效果。通过理论模型处理及对多个处理方法的对比,表明本方法的去噪能力优于二维多级中值滤波器,而保护有效信号的能力要优于经典中值滤波器。实际剖面的处理效果也表明,此方法不仅能有效地滤除脉冲噪声,而且能很好地保护有用信号。
The random noise in seismic exploration is always one of key factors affecting the S/N ratio of seismic signal.Facing on the shortcoming that 1-D classic median filter has strong ability of denoise for seismic signal,which leads the useful information to be destroyed,the paper defines a window length to control threshold based on the feature of noises,then,selects adequate time-variant window to carry out median filtering according to the relationship between seismic data and the threshold, achieving the optimal balancing effect between useful signal and different strength of noises.It is shown by theoretical model processing and correlation of multiple processing methods that the method is superior to 2-D multilevel median filter in the ability of denoise and to the classic median filter in protection of useful signals.The effects of processed practical profiles also showed that the method can not only eliminate pulse noise effectively, but also protect useful signal well.
出处
《石油地球物理勘探》
EI
CSCD
北大核心
2008年第3期327-332,370,24,共8页
Oil Geophysical Prospecting
关键词
随机噪声
时变中值滤波
经典中值滤波
去噪
滤波窗口
random noise
time-variant median filtering
classic median filtering
denoise
filtering window
