Estimation of quality factor Q plays a fundamental role to enhance seismic resolution via absorption compensation of the near surface layer. Two methods have been widely used to estimate the Q value, the spectral ratio method (SRM) and the centroid frequency shift method (CFS). However, the SRM is too sensitive to noise and performance poorly under the low signal-to-noise (SNR) data. When the spectrum of the wavelet deviates from the Gaussian-like, the Q value estimated by the CFS will produce deviation. In this paper, we present a new method for Q estimation. Through establishment the relationship between reference wavelet logarithmic spectrum area and centroid frequency shift of logarithmic spectrum (LSCFS) to inverse Q value. we use the cross-hole data to analyze the stability and reliability of this method. The results of absorption compensation shows that our method apply much more to the real data, which indicates that the LSCFS method is more robust and more precise for Q estimation than the other two methods.
展开▼