An efficient local operator based Q-RTM algorithm with multi-stage optimization

Tong Zhou,Wenyi Hu,Jieyuan Ning

DOI: https://doi.org/10.1190/geo2017-0026.1

IF: 3.264

2018-01-01

Geophysics

Abstract:Most existing Q-compensated reverse time migration (Q-RTM) algorithms are based on pseudo-spectral methods. Because of the global nature of pseudo-spectral operators, these methods are not ideal for efficient parallelization, implying that they may suffer from high computational cost and inefficient memory usage for large scale industrial problems. In this work, we report a novel Q-RTM algorithm - the multi-stage optimized Q-RTM method. This Q-RTM algorithm employs a finite-difference method to compensate both the amplitude and the phase simultaneously by uniquely combining two techniques: 1) negative t method for the amplitude compensation; 2) a multi-stage dispersion optimization technique for the phase correction. To prevent high frequency noise from growing exponentially and ruining the imaging results, we apply a finite impulse response (FIR) low-pass filter using the Kaiser Window. Both the theoretical analyses and numerical experiments demonstrate that this Q-RTM algorithm precisely recovers the de...