基于Contourlet变换的地震噪声衰减

Noise suppression of seismic data based on contourlet transform

在地震勘探中,随机噪声是一种频带较宽的干扰波,常规的去噪方法效果不理想。Contourlet变换具有多尺度和多方向性特点,在计算中,将多尺度分析和方向分析分步进行。首先介绍Contourlet变换的基本原理及实现方法,并进行地震资料仿真计算和实例计算,利用拉普拉斯金字塔塔形分解完成Contourlet变换的多尺度分解,再利用方向滤波器组完成Contourlet各个尺度上的方向分解。在Contourlet变换域对Contourlet变换系数采用阀值处理,再做反变换,从而实现地震数据去噪。结果证明Contourlet变换去噪具有较好的效果,易于实现,有一定的可行性和较好的应用前景。

Random noise is one kind of interference that is broadband and deteriorates signal-to-noise ratio of signals in seismic exploration.Commonly used noise removal methods fail to yield desired results.Contourlet transform is characterized by multi-scale and multi-directivity Noises are eliminated as follows.Firstly,decompose seismic data with Laplacian pyramid,then apply directional filter bank to the resulted data,after that perform threshold process on contourlet coefficients in contourlet domain,and finally carry out inverse transform to accomplish noise removal.In this paper,the author discussed the principles of contourlet transform and its realization,and demonstrated the method with theoretical and read data.