论文部分内容阅读
谱分析是重要的地震信号处理工具。将地震数据变换到频率域是众多处理算法和解释方法的基础。然而,如果地震信号的频率成分随时间发生变化,那么,进行简单的一维(傅里叶)频率变换是不够的。通过滑动时窗(时间较短)可以提高频一时(FT)空间的谱分解的质量,但该方法的缺点是FT分辨率有限。最近提出的以新的小波分析数学理论为基础的变换法能避免这种分辨率的限制,进行质量更高的谱分解。与短时窗傅氏变换相比,可更好地从概念上搞清比例尺—平移平面的连续小波变换。离散小波变换和匹配追逐算法为可以选择的两种小波变换法,对地震记录成图时可将其变换到F—T空间。合成数据和经标定的炸药震源地震数据的F—T分解表明,匹配追逐算法能很好地使谱局部化,并可清地晰地分辨出反射波、直达波、面波和人为噪声。以小波变换为基础的处理算法为改进处理算法和谱解释法提供了新的机会。
Spectral analysis is an important seismic signal processing tool. Transforming seismic data into the frequency domain is the foundation of many processing and interpretation methods. However, if the frequency content of the seismic signal changes over time, then a simple one-dimensional (Fourier) frequency transform is not enough. The quality of spectral decomposition in frequency-only (FT) space can be improved by sliding the time window (shorter time), but the disadvantage of this method is the limited FT resolution. The recently proposed transformation method based on the new mathematical theory of wavelet analysis can avoid this limitation of resolution and perform higher quality spectral decomposition. Compared with the short-time window Fourier transform, the scale-translation plane continuous wavelet transform can be better conceptualized. Discrete Wavelet Transform and Matching Chase Algorithm are two kinds of wavelet transform methods that can be selected. When the seismic record is plotted, it can be transformed into F-T space. F-T decomposition of the synthetic data and the calibrated explosives seismic source seismic data shows that the chasing matching algorithm can well localize the spectrum and clearly distinguish the reflected wave, the direct wave, the surface wave and the man-made noise. Wavelet-based processing algorithms provide new opportunities for improved processing algorithms and spectral interpretation methods.