Dynamic seismic wavelet estimation

determinant wavelet estimation and statistical wavelet estimation are two commonly-used methods at present,both of which have excellence and disadvantages.Conventional convolution model believes that wavelet pattern remain the same during propagating process thus ignores the dynamic feature of wavelet,the two method introduced here are base on the conventional convolution model,but practically proved that wavelet would be absorbed by stratum during propagating process,and the wavelet pattern was fading,which means that the propagation process is actually a dynamically fading process.Stationary convolution model is improved and dynamic convolution model is proposed to estimate the dynamic wavelet.Supposing the reflectance is white noise,the wavelet traveling in the stratum will form an attenuating seismic channel,analyzing the seismic trace in time-frequency,and smooth it with boxcar smoother to obtain wavelet spectrum.Wavelet phase could be obtained through Hilbert transform of wavelet spectrum,and then dynamic wavelet estimation could be accomplished.The feasibility of these methods have been proved by analysis on theoretical models and practical calculations.