Time-Frequency Analysis Method And Its Application In Spectral Decomposition Of Seismic Signals

Time-frequency analysis jointly represents one-dimensional time signal in time and frequency domains,which is one of the most important tools for non-stationary signal analysis.Finding new time-frequency methods to improve the accuracy of timefrequency analysis is one of the research tasks for time-frequency analysis.Non-stationary is the basic characteristic of signals in nature.For example,seismic signals in seismic exploration,machine vibration signals in fault diagnosis,radar signals,etc.In this paper,Chebyshev orthogonal polynomials are firstly used as the basis to asymptotically approximate the input signal.The coefficients of the above approximation are generalized to time-dependent,which are defined as the time-frequency distribution of the input signal and are computed by the discrete cosine transform.Numerical experiments verify that this method can obtain the time-frequency characteristics of the input signals.Compared with the conventional time-frequency analysis methods reveals that the method has good accuracy.Synchrosqueezing wavelet transform is a highprecision time-frequency analysis method proposed in recent years.It reallocates the wavelet transform values to its instantaneous frequencies locations and therefore generates a concentrated time-frequency map.In this paper,the synchrosqueezing shorttime Fourier transform is deduced for the short-time Fourier transform and synchrosqueezing method.Experiments on Synthetic examples and real data verify that the compression of short-time Fourier transform effectively improves the accuracy of time-frequency analysis.The low-frequency anomalies of seismic signals may be attributed to the abnormal attenuation of the high-frequencies in hydrocarbon areas,and low-frequency anomalies are often associated with oil and gas reservoirs.In this paper,the synchrosqueezing short-time Fourier transform is applied to the low-frequency anomalies detection,thereby improving the accuracy of reservoirs prediction.