基于多小波预测方法提高地震数据的信噪比

Improving the Signal-noise Ratio of Seismic Data by Orthogonal Multiwavelets Transformation and w—x Prediction Method

基于二维小波变换把一个二元的地震数据变换为有关时间,空间,频率和波数的局部信息的特点,讨论了用正交多小波变换和w-x预测方法对地震数据进行去噪处理.先用阈值的方法做初步处理,再用预测方法进一步压制噪声,达到较好地提高地震数据信噪比的目的.数值试验表明该方法是有效的,能有效地消除随机干扰,具有很好的空间和时间自适应性.

Two dimension time, space, frequency and wavelet transform changes a binary seismic data into a function of wave number, supplying the local information of the seismic data in four aspects. Base on those, improving the S/N ratio of a seismic data is discussed by orthogonal multiwavelet transform and w- x prediction filter method. First, elementary process is done by wavelet thresholding, then suppressing noise of seismic record has been realized by applying w -- x prediction filte method. Experimental results and practical application shows this method has quite good effect, better self-adaptability in space and time, and better denoising effect.