小波方差与小波熵在信号特征提取中的应用

讨论了单一尺度下的小波方差,并结合信息论中信息熵的定义和物理意义,进一步引进了多尺度下的小波熵;分别以频率突变和幅度突变两种仿真信号为对象,分析了小波方差与小波熵在反映随机信号统计特征方面的特点;最后,以一类钻井信号为例,分别利用两种方法分析和提取了信号在强噪声环境下的脉冲特征。仿真和实例说明,相对于小波方差对尺度选择的依赖,小波熵可以综合各尺度的信息,能够从整体上更有效地提取信号特征。

The Application of Wavelet Variance and Wavelet Entropy in Signal Feature Extraction

The wavelet variance under the single scale is discussed,and in combining the information entropy in information theory with physical sense,the wavelet entropy under the multi-scale is introduced in this paper.With two kinds of simulation signals of frequency mutation and amplitude mutation as the objective,the characteristics of wavelet variance and wavelet entropy in reflecting the statistic features of random signals are analyzed.Finally taking the signals from the drilling as an example,two kinds of methods are used to analyze and extract the signal pulses features under the strong noise environs.Accordingly,the simulation and actual examples indicate that being dependent on the wavelet variance relative to the scale selection,wavelet entropy can extract the signal features wholly and effectively with the information of integrated scale.