一种基于小波分析的地震信号特征向量提取方法

A Method Based on Wavelet Analysis Seismic Signal Feature Vector Extraction Method

由于地震信号的复杂性,被引入的信号可能是噪声,但也有可能带有与地质结构相关的重要信息,前者应该被消除,后者则应该保留,如何对信号类型进行判断成为了地震信号处理中的难点。文中假定被引入信号属于不同的信号类型,提出一种基于小波分析的地震信号特征向量提取方法,对被引入的信号计算其在各个频率尺度下包含的无偏估计量,取得该信号在这个尺度下的无偏估计量构成该信号的特征向量,然后根据特征向量值计算该信号的模式分类,从而判断该信号类型,对不同信号类型采用不同处理方式,仿真实验证明了本方法的有效性。

Because of the complexity of seismic signal, the introduced signal may be noise. But it may be the potential and geological structure associated with important information too.the former should be eliminated, the latter should be retained.The judgment of signal type has become the difficult point in seismic signal processing. This paper assumes that the introduced signals are belonging to different types of signal. The method is proposed based on wavelet analysis of the seismic signal feature extraction, the introduced signal is calculated in various frequency scales . The unbiased estimation is as feature vector, and then according to the feature vector to calculate the signal pattern classification, so as to judge the signal types.