The method research of stripping strong reflection of coal seams

The coal seams have the characteristics of low speed, low density. The existence of coal seams results in the strong reflection on the seismic profile, which makes the reflection information of oil (gas) bed difficult to identify and increases the difficulty for reservoir prediction. In this paper, we introduce two strong reflection stripping methods, including Multi-wavelets decomposition and reconstruction, Matching Pursuit controlled by horizon and seismic wavelet. The method of Multi-wavelets decomposition and reconstruction is to construct wavelet library by solving a covariance matrix, then choose the particular frequency wavelet to reconstruct the targeted event. The original Matching Pursuit is a greedy algorithm which needs large amount of calculation, but the Matching Pursuit controlled by horizon and wavelet we proposed adds the constraint of horizon and seismic wavelet during the construction of Sparse Dictionary, which can improve the efficiency of calculating the signal support set of atoms and reconstruct the strong reflection event exactly. Meanwhile, this paper compares the effect of these two methods. And we can see that the Matching Pursuit controlled by horizons and seismic wavelet has better location accuracy for strong reflection, and Multi-wavelets decomposition and reconstruction method may cause event 'disconnected'.