In this paper,we present a new method for seismic stratigraphic absorption compensation based on the adaptive molecular decomposition.Using this method,we can remove most of the effects resulting from wavelets truncation and interference which usually exist in the common time-frequency absorption compensation method.Based on the assumption that the amplitude spectrum of the source wavelet is smooth,we first construct a set of adaptive Gabor frames based on the time-variant properties of the seismic signal to transform the signal into the time-frequency domain and then extract the slowly varying component（the wavelet＇s time-varying amplitude spectrum） in each window in the timefrequency domain.Then we invert the absorption compensation filter parameters with an objective function defined using the correlation coefficients in each window to get the corresponding compensation filters.Finally,we use these filters to compensate the timefrequency spectrum in each window and then transform the time-frequency spectrum to the time domain to obtain the absorption-compensated signal.By using adaptive molecular decomposition,this method can adapt to isolated and overlapped seismic signals from the complex layers in the inhomogeneous viscoelastic medium.The viability of the method is verified by synthetic and real data sets.
