Reservoir prediction method based on synchrosqueezing compression wavelet transform

Seismic signals are non-stational, and spectral decomposition is an important method for studying the properties of non-stationary signals. Conventional spectrum decomposition methods can not simultaneously have high time-frequency resolution, which can not satisfy the requirement of high-precision seismic data interpretation at present. Therefore, the text proposes a new spectral decomposition method with high resolution in time and frequency domain which named synchrosqueezing wavelet transform (SSWT). Since the wavelet coefficients of the wavelet transform are compressed and rearranged only in the frequency axis of SSWT, it can reconstruct the signals. Based on the basic principle of the SSWT, the relevant properties such as anti-noise ability and invertibility were analyzed, the frequency modified synchrosqueezing wavelet transform algorithm was used to effectively reduce the time-frequency spectrum ambiguity when the variation rate of instantaneous frequency of the signal not equals to zero. Through the spectral decomposition of the actual seismic data by using continuous wavelet transform (CWT), three parameter wavelet transform (TP), S transform (ST)and SSWT, the superiority of the SSWT with higher time-frequency resolution was verified.