Abstract: In anisotropic media, seismic waves have different velocities in different propagation directions. The use of isotropic wave equations cannot accurately describe the propagation of seismic waves, which can lead to inaccurate positioning of deep events during imaging. On the basis of the TTI medium elastic wave equation, this paper sets the shear wave velocity to a smaller value and introduces a source loop to eliminate pseudo shear wave artifacts. On the other hand, logging data is used to constrain and correct the parameters field, in order to improve the imaging accuracy of deep and complex structures. The reverse time migration(RTM) method can have an impact on imaging due to low-frequency noise. Therefore, based on the Laplacian filtering operator for denoising, the paper extracts the angle domain common imaging point gathers(ADCIGs) and uses angle-selected stacking imaging to suppress strong amplitude low-frequency noise caused by inaccurate migration parameters. The real data requires a large amount of computation during imaging. In this article, the CPU/GPU cooperative operations are used to accelerate migration and improve computational efficiency. The model and real data results show that the TTI-RTM method based on logging data constraints can accurately image deep and complex structural target areas, which also proves the feasibility of the method.

Key words: TTI media, RTM, pseudo elastic wave, logging data, ADCIGs, angle-selected stack imaging