Continuous time-varying Q-factor estimation method in the time-frequency domain

Conclusion

We define the local centroid frequency using shaping regularization in the time-frequency domain and discuss the theoretical basis of the local centroid frequency in detail in this paper. The local centroid frequency is calculated locally using information around a certain time so that a reasonable result still can be obtained at the time without information. By combining the local centroid frequency with the CFS method, we propose a continuous time-varying Q-estimation method, namely, the LCFS method. The theoretical model shows that the proposed method can obtain reasonable time-varying Q-factors even in an environment with random noise. Compared with the traditional methods, the proposed method avoids picking the target layer and can still obtain a stable estimation result in the absence of spectrum information. At the same time, experiments with synthetic examples and field data confirm that the inverse Q filtering method based on continuous time-varying Q-factors can better recover the phase and amplitude of the attenuation signal and improve the resolution of seismic data.

Continuous time-varying Q-factor estimation method in the time-frequency domain