地震信号分数域局部功率谱估计及应用

Seismic signal fractional local power spectrum estimation

将谱估计技术与分数阶傅里叶变换(FrFT)结合,提出了一种新的分数域局部功率谱估计方法。首先运用最大峰度系数原理找到分数阶傅里叶变换的最优阶次,求取该阶次下分数域功率谱的分数域二维分布,利用分数阶傅里叶变换的旋转性质得到分数域功率谱的时频表达式,同时利用自回归模型(AR)谱估计技术给出了分数域功率谱时频分布的算法。提出的两种谱估计算法时频特性均优于二维谱图。理论模型和实际资料的测试验证了本文方法的有效性。

We propose in this paper a new method of fractional local power spectrum estimation by combining power spectrum estimation and fractional Fourier transform (FrFT). First, the maximum kurtosis in the fractional domain is found out to determine the optimal order for fractional power spectrum. Then the fractional 2-D distribution of fractional power spectrum in the optimal order is calculated. The time-frequency 2-D distribution of fractional power spectrum can be obtained with the rotation of fractional Fourier transform. Similarly, auto-regressive (AR) spectrum estimation can be used to calculate the time-frequency 2-D distribution of fractional power spectrum. The time-frequency properties obtained by these two methods are superior to that of the classic spectrogram. Tests on theoretical model and seismic signals show that the proposed method is efficient. © 2016, Editorial Department OIL GEOPHYSICAL PROSPECTING. All right reserved.