A Dictionary Learning Approach for Interval Q Estimation and Compensation

A seismic trace corresponding to an anelastic layered earth model can be sparsely represented by a structured dictionary of properly attenuated wavelets, through a non-stationary sparse deconvolution. The sparseness of the coefficients (reflectivity series), however, considerably decreases when the wavelets are incorrectly attenuated. We propose an adaptive parametric dictionary learning strategy for the interval-Q estimation and compensation by adaptively training the dictionary from the trace to sparsely represent it.

We assume a piecewise Q-model by dividing the dictionary elements into several groups, each comprises a number of wavelets whose temporal supports are close to each other and can be described by a single Q value. The dictionary is learnt iteratively where at each iteration only one group of the wavelets are optimized by searching for the corresponding optimum Q value, leading to an iterative construction of the Q-model. The proposed method is tested on both synthetic and field data and the results obtained demonstrate the stability and accuracy of the method for interval-Q estimation and compensation.