基于二维小波变换的随机噪声压制方法研究

Method of 2-D wavelet transform in attenuating random noise

1. 中国石化石油勘探开发研究院,北京 100083;2. 中国地质大学,北京 100083;3. 中国科学院地质与地球物理研究所,北京 100029

Exploration and Production ResearchInstitute, SINOPEC, Beijing 100083, China

在地震勘探中,随机噪声是一种频带较宽、严重影响有效波的干扰波,常用的一维去噪方法效果不理想。小波变换是一种时频分析方法,根据它的分频和局部分析能力,能有效地消除随机干扰,保留有效波的中、高频成分,经过小波重构,可恢复有效波信号。针对地震信号随机干扰的特点,运用二维小波变换的理论,设计了相应的变换域去噪滤波器。此方法的特点是计算速度快,稳定性好,自适应性强,能对各种地震数据进行去噪处理,模型数据与实际数据的应用效果证明,二维小波变换具有较强的信噪分离能力。

The random noise is an interference wave that has broadband and deteriorates signal-to-noise ratio of signal in seismic exploration. The commonused 1-D noise-eliminated methods have poor effects. The wavelet transform is a time-frequency analysis method, which can effectively eliminate random interference and preserve high and middle frequency components of signal according to its frequency-shared and local analyzed ability. The rationale of wavelet transform de-noising is introduced. A filter is designed to attenuating noise in accordance with the characteristics of seismic signals having a random disturbance with the theory of the 2-D wavelet transform. It is characterized by fast computation, good stability, desirable adaptive ability, and applications to various seismic data. Trial computations of a theoretical record and an actual field record showed that 2-D wavelet transform has obvious effect on separation of signal from noise.