小波包F-X域前后向预测去噪

Noise Suppression through Wavelet Package Transformation and F-X Domian Prediction

根据地震资料不同的频段信噪比不同应采取分频去噪的原理，提出了地震信号小波包分解与F－X城预测联合去噪技术，并给出了具体实现的方法和步骤。最后展示新疆焉耆博湖地区实际资料的处理效果。处理表明，该方法在提高信噪比的同时，又能保持分辨率。

According to the principle that seimic data withdifferent signal- noise ratio in each frequency bandshould separatly suppress noise,this paper presents a newnoise suppression method which combines wavelet packagedecomposition with F - X domain prediction,and explainsthe detail steps. The actual data processing results showthat this method can impove both signal- noise ratio andresolution.