小波包多阈值法在地震信号去噪中的应用研究

Application of wavelet packet multi-threshold method in seismic signal denoising

地震信号中通常含有各种干扰噪声,严重影响了地震资料的信噪比和分辨率,小波包变换是地震资料去噪的有效方法之一。针对传统小波包阈值去噪不明显和存在失真的问题,提出一种基于多阈值函数的小波包地震信号去噪方法。对地震波信号进行小波包分解,并对小波包分解系数按照频率大小的顺序进行排列,根据分解的系数处于不同频带选取不同的阈值准则进行去噪处理,对得到的系数进行重构,可有效地去除地震信号中的噪声。对仿真地震信号以及实际地震信号进行小波包多阈值去噪处理,实验结果表明,该方法较好地去除了干扰噪声保留了有用信号,去噪效果明显且失真小,有效地提高了地震资料的分辨率。

Since the seismic signal generally contains a variety of interference noises, which affects the signal-to-noise ratio and resolution of seismic data seriously, and because the wavelet packet transform is an effective way of the seismic signal de- noising, a seismic signal denoising method based on wavelet packet of muhi-threshold function is proposed for the traditional wavelet packet threshold denoising method has unobvious denoising effect and distortion. The wavelet packet decomposition is conducted for seismic wave signal, and the wavelet packet decomposition coefficients are arranged according to the frequency real values. Different threshold criterions are selected to conduct denoising processing according to the decomposed coefficients in dif- ferent bands to reconstruct the obtained coefficients, which can eliminate the noise in seismic signal effectively. The denoising processing of wavelet packet multi-threshold for the simulation seismic signal and practical seismic signal were conducted. The experimental results show this method can eliminate the interference noise availably while reserving the useful signal, has good denoising effect and low distortion, and can improve the resolution of seismic data efficiently.